Anda di halaman 1dari 121

AP-R422-12

AUSTROADS RESEARCH REPORT

Effectiveness of Road Safety Engineering Treatments

Effectiveness of Road Safety Engineering Treatments

Effectiveness of Road Safety Engineering Treatments Published November 2012

Austroads Ltd 2012 This work is copyright. Apart from any use as permitted under the Copyright Act 1968, no part may be reproduced by any process without the prior written permission of Austroads.

Effectiveness of Road Safety Engineering Treatments ISBN 978-1-921991-60-8

Austroads Project No. ST1571 Austroads Publication No. AP-R422-12

Project Manager Maurice Cammack, Main Roads WA

Prepared by Blair Turner, Lisa Steinmetz, Adrian Lim, Karen Walsh ARRB Group

Published by Austroads Ltd Level 9, Robell House 287 Elizabeth Street Sydney NSW 2000 Australia Phone: +61 2 9264 7088 Fax: +61 2 9264 1657 Email: austroads@austroads.com.au www.austroads.com.au

Austroads believes this publication to be correct at the time of printing and does not accept responsibility for any consequences arising from the use of information herein. Readers should rely on their own skill and judgement to apply information to particular issues.

Effectiveness of Road Safety Engineering Treatments

Sydney 2012

About Austroads Austroads purpose is to:


promote improved Australian and New Zealand transport outcomes provide expert technical input to national policy development on road and road transport issues promote improved practice and capability by road agencies. promote consistency in road and road agency operations.

Austroads membership comprises the six state and two territory road transport and traffic authorities, the Commonwealth Department of Infrastructure and Transport, the Australian Local Government Association, and NZ Transport Agency. Austroads is governed by a Board consisting of the chief executive officer (or an alternative senior executive officer) of each of its eleven member organisations:

Roads and Maritime Services New South Wales Roads Corporation Victoria Department of Transport and Main Roads Queensland Main Roads Western Australia Department of Planning, Transport and Infrastructure South Australia Department of Infrastructure, Energy and Resources Tasmania Department of Transport Northern Territory Department of Territory and Municipal Services Australian Capital Territory Commonwealth Department of Infrastructure and Transport Australian Local Government Association New Zealand Transport Agency.

The success of Austroads is derived from the collaboration of member organisations and others in the road industry. It aims to be the Australasian leader in providing high quality information, advice and fostering research in the road transport sector.

Effectiveness of Road Safety Engineering Treatments

CONTENTS
1 2 2.1 2.2 2.3 2.4 3 4 INTRODUCTION ................................................................................................................... 1 METHOD ............................................................................................................................... 2 Identifying Gaps in Knowledge ............................................................................................... 2 Filling the Gaps Knowledge ................................................................................................... 2 Selected Issues ..................................................................................................................... 2 Determining Treatment Effectiveness .................................................................................... 3 REVIEW OF LITERATURE ON CRASH TREATMENT EFFECTIVENESS ........................... 5 DISCUSSION AND RECOMMENDATIONS ........................................................................ 14 PRIORITISING GAPS IN KNOWLEDGE ................................................. 16 REVISIONS TO TREATMENT EFFECTIVENESS FIGURES................... 19 SUMMARY OF RESEARCH FOR PRIORITY TREATMENTS ................. 25

REFERENCES ............................................................................................................................. 15 APPENDIX A APPENDIX B APPENDIX C C.1 C.2 C.3

Guide Posts ......................................................................................................................... 26 Chevron Alignment Markers................................................................................................. 29 Pavement Markings ............................................................................................................. 31 C.3.1 Provision of Edge Lines.......................................................................................... 31 C.3.2 Provision of Centrelines ......................................................................................... 33 C.3.3 Provision of Centreline and Edge Line ................................................................... 34 C.4 Profile Line Marking ............................................................................................................. 37 C.4.1 Profile Edge Lining, Edge Line (Shoulder) Rumble Strip, or Shoulder Grooving ..... 37 C.4.2 Profile Centre Lining, Centreline Rumble Strip ....................................................... 40 C.5 Signs Regulatory .............................................................................................................. 43 C.5.1 General: Regulatory Signs at Intersections ............................................................ 45 C.5.2 Stop Signs at Three-leg Intersections..................................................................... 45 C.5.3 Stop Signs at Four-leg Intersections....................................................................... 46 C.5.4 Four-way Stop Signs .............................................................................................. 46 C.5.5 Give-way Signs ...................................................................................................... 46 C.5.6 Stop Sign to Give-way Sign .................................................................................... 46 C.5.7 Give-way Sign to Stop Sign .................................................................................... 47 C.5.8 U-turn and Right-turn Ban ...................................................................................... 47 C.6 Traffic Signals ...................................................................................................................... 48 C.6.1 New Signals ........................................................................................................... 48 C.6.2 Effect of Turn Phases ............................................................................................. 51 C.7 Signal Visibility ..................................................................................................................... 56 C.8 Channelisation at Intersections Splitter and Median Islands.............................................. 59 C.8.1 General .................................................................................................................. 60 C.8.2 Splitter Islands........................................................................................................ 60 C.8.3 Median Islands ....................................................................................................... 60 C.9 Grade Separated Intersections ............................................................................................ 62 C.10 Right-turn Lane Provision..................................................................................................... 64 C.10.1 All Locations ........................................................................................................... 69 C.10.2 Signals ................................................................................................................... 69 C.10.3 Unsignalised Intersections ..................................................................................... 69 C.10.4 Rural ...................................................................................................................... 70

Austroads 2012 i

Effectiveness of Road Safety Engineering Treatments

C.11 C.12 C.13 C.14 C.15 C.16 C.17

C.18

C.19 C.20 C.21

C.10.5 Urban ..................................................................................................................... 70 C.10.6 Painted and Physical .............................................................................................. 71 C.10.7 X and T-intersections ............................................................................................. 71 Extend Right-turn Lane ........................................................................................................ 74 Left-turn Lane Provision ....................................................................................................... 76 Lane Width in Tunnels ......................................................................................................... 79 Overtaking Lanes................................................................................................................. 80 Superelevation ..................................................................................................................... 83 Impact Attenuators ............................................................................................................... 86 Signs Advisory .................................................................................................................. 89 C.17.1 Curve Warning Signs ............................................................................................. 89 C.17.2 Speed Advisory Signs ............................................................................................ 90 Pedestrian Treatments......................................................................................................... 92 C.18.1 Pedestrian Fencing and Barriers ............................................................................ 92 C.18.2 Improved Lighting................................................................................................... 92 C.18.3 Marked Crossings .................................................................................................. 93 C.18.4 Pedestrian Overpasses .......................................................................................... 94 C.18.5 Pedestrian Signals ................................................................................................. 95 C.18.6 Refuges.................................................................................................................. 96 C.18.7 Roundabouts .......................................................................................................... 97 C.18.8 Raised Pedestrian/Wombat Crossings ................................................................... 97 Transverse Rumble Strips.................................................................................................. 101 Street Closure .................................................................................................................... 103 Street Lighting.................................................................................................................... 105 C.21.1 New Lighting All Sites........................................................................................ 109 C.21.2 New Lighting Intersections ................................................................................ 109 C.21.3 New Lighting Mid-blocks ................................................................................... 109 C.21.4 New Lighting Rural ............................................................................................ 109 C.21.5 New Lighting Rural Intersections ....................................................................... 109 C.21.6 New Lighting Urban ........................................................................................... 109 C.21.7 New Lighting Urban Intersections ...................................................................... 109 C.21.8 New Lighting Motorway Freeway Interchanges ................................................. 109 C.21.9 Railway Crossings................................................................................................ 109 C.21.10 Upgrade Existing Lighting .................................................................................... 110 C.21.11 Effect on Different Crash Types and Severities .................................................... 110 C.21.12 Time of Day.......................................................................................................... 110

Austroads 2012 ii

Effectiveness of Road Safety Engineering Treatments

TABLES
Table 3.1: Summary of treatment effectiveness ........................................................................ 6

Austroads 2012 iii

Effectiveness of Road Safety Engineering Treatments

SUMMARY
A reliable knowledge base about the effectiveness of different road safety engineering treatments allows informed decisions for the development of future road safety programs. This information is used to allocate resources to the most cost-effective projects (i.e. ones that will reduce casualties the most per dollar spent). Inaccurate information on treatment effect may lead to inefficient use of limited resources. The current project aims to address this issue and to provide updated advice on the road safety benefit of engineering treatments, based on literature reviews. Previous Austroads research has identified that there is a lack of reliable information regarding the effectiveness of different road safety engineering treatments. During the first stage of this project, a number of treatments or road safety features for which the knowledge base was inadequate were identified. The assessment was based on literature reviews as well as stakeholder priorities. Given the large number of gaps, and the time and budget available for this project, the knowledge gaps were prioritised. A summary of the crash reduction effects for each of these treatments is provided, along with an assessment of the level of confidence in the figure (based on factors such as consistency of results from various studies and number of studies). Where possible, information on the crash reduction for different severity outcomes and crash types is provided. However, information on these factors is relatively scarce. The project has identified crash effectiveness for 57 treatment types and 126 crash effectiveness values have been derived for these. Compared to an earlier study on this topic (Austroads 2010a), there is now more information available on treatment effectiveness (126 values compared to 104). The project has also led to an increased level of confidence in knowledge about treatments; over half of the values are now allocated a medium or high level of confidence. Although this project has improved the reliability of information on the effectiveness of treatments, a number of knowledge gaps remain that can be filled by further reviews of literature and experimentation. During the course of this project, links were developed with an OECD initiative to address crash reduction effectiveness of different treatments. This initiative has led to the establishment of an international collaboration to improve information on the crash reduction effectiveness of treatments. There is great merit in Australian and New Zealand road safety professionals contributing to the international collaboration once key gaps have been identified and prioritised by the international OECD group.

ACKNOWLEDGEMENTS
The authors would also like to acknowledge the input of the project steering group. This group is made up of representatives from Australian road agencies, the NZ Transport Agency, the Federal Department of Infrastructure and Transport and AAA/AusRAP. In addition, the authors would also like to acknowledge the input of the Safety Task Force who provided comments on the document.

Austroads 2012 iv

Effectiveness of Road Safety Engineering Treatments

INTRODUCTION

Previous Austroads research (Austroads 2010a) identified that there was a lack of reliable information regarding the effectiveness of different road safety engineering treatments. The effectiveness relates to the expected reduction in crashes from the use of treatments in different circumstances. The effectiveness of a treatment can be expressed as a Crash Modification Factor (CMF). This represents the relative change in crash frequency due to a specific change in the road or its immediate environment (Austroads 2012, p49). Effectiveness in Australia and New Zealand has traditionally been presented using Crash Reduction Factors (CRFs), which presents the expected percentage reduction in crashes. The term CMF is now used more widely overseas, although both terms are used in this current report 1. This information is used to allocate resources to the most cost-effective projects (i.e. ones that will reduce casualties the most per dollar spent). Inaccurate information on treatment effect may lead to inefficient use of limited resources. This work is one component of a large program of research. The Austroads National Risk Assessment Model project has three main objectives. The first is the development of a risk assessment model, intended to be used nationally as a way of identifying crash risk. The second task involves the development of a national program for risk assessment with associated guidelines. A third task involves improving information on the crash reduction effectiveness of various road safety treatments. This included the development of a stand-alone report on a standard approach for evaluation of treatment effectiveness for adoption by road authorities. This work has been published as An Introductory Guide for Evaluating Effectiveness of Road Safety Treatments (Austroads 2012). An associated task, and the subject of this report, involves improving knowledge on the road safety benefit of engineering treatments. During the first stage of this project, treatments or road safety features were identified for which there were significant gaps in knowledge in terms of crash reduction effectiveness. The work also identified a method for prioritising each of these issues. That process is outlined in Section 2 of this report. Section 3 provides information on local and international literature on each of the priority issues identified. For completeness, the section also provides information from previous work on this topic (Austroads 2010a) so that the sum of current knowledge on crash reduction effectiveness for engineering treatments is provided in this report. The final section (Section 4) provides concluding comments and recommendations.

A Crash Reduction Factor indicates the expected percentage reduction in crashes following the introduction of a treatment. A Crash Modification Factor is a proportion that represents the relative change in crash frequency due to a specific change in the road or its immediate environment (Austroads 2012).

Austroads 2012 1

Effectiveness of Road Safety Engineering Treatments

2
2.1

METHOD
Identifying Gaps in Knowledge

During the first stage of this project, a number of treatments or road safety features were identified where knowledge of effectiveness in reducing crashes was not adequate. This assessment was based on literature reviews (Austroads 2010a, 2010b) as well as stakeholder priorities (Austroads 2009) 2. For each of these issues it was identified that there was a lack of sufficient research on the crash reduction effect, or that the research which did exist was not considered of suitable quality. The identified gaps in knowledge were then prioritised. Appendix A provides the results of this assessment.

2.2

Filling the Gaps Knowledge

Two main options were identified to address the gaps in knowledge identified above. The most cost-effective method to address gaps in knowledge is typically through review of previously published research. Often research will have been conducted somewhere in the world that either addresses a gap in knowledge, or goes part way in addressing this gap. New studies are conducted every year, and the knowledge base is continually being updated and expanded. Although extensive literature reviews have been conducted by Austroads on many issues relating to the effectiveness of road safety engineering treatments, there is still a need to assess any new knowledge which has been generated. It was recommended that any future research to fill gaps in knowledge reassess the available literature before embarking on empirical research. If a gap in knowledge is deemed important enough to require further research, and adequate information cannot be obtained from existing literature, then a trial of the treatment will typically be required. Austroads (2012) provides guidance on methods for evaluating the effectiveness of road safety treatments. Given the cost associated with trials of treatments, and the timeframe required (three to five years of data is typically required after the treatment is installed), this current project focuses on literature review to fill gaps in knowledge.

2.3

Selected Issues

Based on the existing gaps in knowledge, and the prioritisation of the tasks, the treatments listed below were assessed:

In 2010/11: grade separation of intersections guide posts lane width in tunnels overtaking lanes pavement markings edge line pavement markings centreline pedestrian treatments

Engineering research priorities were identified in a previous project (Austroads 2009) via a survey of Austroads Road Design Review Panel and Traffic Management Review Panel members.

Austroads 2012 2

Effectiveness of Road Safety Engineering Treatments

signs advisory street lighting superelevation signs regulatory delineation continuity lines, marking in wet conditions channelisation at intersections splitter and median islands impact attenuators intersection right-turn lane provision intersection extend right-turn lane intersection left-turn lane provision intersection signal visibility line marking profile edge lines street closure traffic signals.

In 2011/12:

It should be noted that some other treatments were identified that had a higher priority than the ones identified above (e.g. clear zones and vegetation). However, since extensive research is currently being conducted on them within Australasia, these topics were not reviewed as part of this project. The crash effectiveness for these will be updated once this work is complete. The treatments assessed as part of this project (listed above) complimented previous work which summarised research from 2004 to 2009 (Austroads 2010a, Road Safety Engineering Risk Assessment: Part 6: Crash Reduction Factors).

2.4

Determining Treatment Effectiveness

A literature review was conducted for each of the topics listed above. Given the large number of topics, and the limited time and budget available for this task, each review was limited in scope and could not be considered exhaustive. The main focus was on recent research since 1999, and on research conducted in Australia and New Zealand. However, other key references have also been assessed and included where relevant. Only research that contained information about the crash reduction benefit from treatments was included in the review, and this greatly reduced the number of relevant publications. Much of the research presented crash reductions for all crash severities (including property damage only). In many cases, treatment effectiveness for casualty crashes was also available. Where information was available, effectiveness values were based on casualty crashes. Often for any one issue, a variety of treatment effectiveness values were identified. In general, an average crash reduction value was taken across all studies that were considered methodologically robust. It would have been preferable to use a weighted average based on robustness (such procedures exist), however due to limited budget a more simplistic approach was taken. Outliers were often excluded, however they were considered on a case by case basis. In some cases, logic checks were used to help refine a figure.

Austroads 2012 3

Effectiveness of Road Safety Engineering Treatments

Given the limited budget available for this work, a qualitative indication on the level of confidence for each of the factors provided was estimated. This was based on the number of studies that provide crash reduction information, the consistency of the results, methodological robustness, the age of the research, and the country in which the research was conducted. For instance, where four or more studies were available, each of which provided reasonably consistent results and had robust methodologies, a high level of confidence was assumed. This same approach was used in Austroads (2010a). Section 3 provides a summary of the results from a review of literature on each of the topics listed as high priority.

Austroads 2012 4

Effectiveness of Road Safety Engineering Treatments

REVIEW OF LITERATURE ON CRASH TREATMENT EFFECTIVENESS

Table 3.1 presents a consolidated summary of treatment effectiveness for a range of treatments, expressed as Crash Reduction Factors (CRFs) and Crash Modification Factors (CMFs). Each factor has been rounded (to the nearest 5%). The results have been collated from previous work (Austroads 2010a, which summarised research from 2004 to 2009) as well as this current project (2010/11 and 2011/12). Appendix B provides a summary of treatment effectiveness values that have been updated since the completion of Road Safety Engineering Risk Assessment: Part 6: Crash Reduction Factors (Austroads 2010a). Details of the research conducted in 2010/11 and 2011/12, including the literature assessed can be found in Appendix C. Details of research from earlier work (up to 2008/09) can be found in the appendices to Austroads (2010a). It is recognised that the effectiveness of treatments on different levels of severity is also of interest. This is particularly relevant in the Safe System context where there is a focus on reduction in fatal and serious injury crash outcomes. There is also interest in quantifying the effect of treatments on different crash types (e.g. head-on). Where this information is available, this is also presented in Appendix C, and in some cases, where the information is robust enough, also in Table 3.1. However, much of the research does not provide a breakdown by severity or crash type, and where information does exist it is often based on very small sample sizes and so is quite variable. In addition, some research (e.g. Hauer 2006 & 2009) indicates that it is not technically possible to accurately determine changes in severity due to the frequency-severity indeterminacy. This states that due to under-reporting of crashes it is not possible to determine whether the frequency of crashes has reduced, or whether there has been a transference to higher or lower levels of severity. Table 3.1 also provides a measure of confidence in the factors provided. As discussed in Section 2.3, this is a qualitative measure that is based on the number of studies that provide crash reduction information, the consistency of the results, methodological robustness, the age of the research, and the country in which the research was conducted. An indication of the year in which a treatment type was assessed is also provided in Table 3.1. This is of interest as the amount of research on crash effectiveness has increased in recent years. In some cases, new research merely validates that of older studies on the same topic. In other cases, the new research can significantly change the expected crash reduction effectiveness (especially when there was previously a high degree of uncertainty). This information will also assist with planning and prioritisation of tasks for investigation in the future. Note that the crash reduction figures provided for pedestrian treatments are given for changes in pedestrian casualties only. Little information exists for these treatments on changes to all casualty crashes. Similarly, reductions for street lighting relate only to changes in night-time casualties. Some information does exist on the effect of street lighting at all times of the day, but this was not comprehensive enough to produce a separate factor. Although this report is primarily concerned with road safety engineering treatments, some information is provided on the safety implications of changing road design elements (e.g. superelevation). Further information on the safety implications of road design can be found in a separate report (Austroads 2010b).

Austroads 2012 5

Effectiveness of Road Safety Engineering Treatments

For a number of treatments that were reviewed, treatment effectiveness was unable to be determined. These included:

Delineation treatments: continuity lines, provision of wide edge lines, marking in wet conditions, provision of word and symbol pavement markings, transverse rumble strips (at curves, railway level crossings and intersection approaches). Intersection treatments: extending right-turn lanes. Pedestrian treatments: installation of pedestrian signals, installation of a marked pedestrian crossing, converting marked pedestrian crossing to signals. Traffic management treatments: closing a street.

The review identified some evidence for a number of additional treatments. Although this information is of interest (and summarised in Appendix C), there was insufficient evidence to determine the extent of treatment effectiveness (e.g. provision of edge lines at curves, addition of yellow reflective tape to signal heads, and channelisation in rural environments). In such cases, the relevant appendix often includes a discussion indicating that safety benefits are likely to be associated with such treatments, although there was insufficient evidence to gauge treatment effectiveness during this project.
Table 3.1: Summary of treatment effectiveness
Treatment Environment type Crash reduction factor Crash modification factor 0.95 0.95 Confidence Year most recently assessed 2010/11 2008/09 Reference location

Delineation treatments Install guide posts Install Raised Reflective Pavement Markers (RRPMs) Install chevron alignment markers Pavement markings All All 5% 5% Low Medium Appendix C.1 Austroads 2010a

Horizontal curves

25%

0.75

Low

2010/11

Appendix C.2

Provide edge line Provide centreline Provide combined edge and centreline Provide painted speed limits

10% 20% 30%

0.9 0.8 0.7

Low Low Low

2010/11 2010/11 2010/11

Appendix C.3.1 Appendix C.3.2 Appendix C.3.3

0% 20% (all) 40% (run-off-road) 15% (all) 30% (head-on)

1.0 0.80 (all) 0.6 (run-off-road) 0.85 (all) 0.70 (head-on)

Low Medium Low Medium Low

2008/09 2011/12

Austroads 2010a Appendix C.4.1

Profile line marking

Provide profile edge lines Provide profile centreline

2011/12

Appendix C.4.2

Austroads 2012 6

Effectiveness of Road Safety Engineering Treatments

Treatment

Environment type

Crash reduction factor

Crash modification factor 0.85 0.7 0.40

Confidence

Year most recently assessed 2011/12 2011/12 2011/12

Reference location

Intersection treatments Signs regulatory Install stop sign at T-intersection Install stop sign at X-intersection Install four-way stop sign at X-intersection Install give-way sign all intersections Install right-turn ban, or U-turn and right-turn ban Traffic signals Install traffic signals Provision of fully controlled rightturns 15% 30% 60% Low Medium Low Appendix C.5.2 Appendix C.5.3 Appendix C.5.4

25%

0.75

Low

2011/12

Appendix C.5.5

60%

0.40

Medium

2011/12

Appendix C.5.8

30% 35% (all casualty) 60% (right through) 45% (adjacent direction)

0.7 0.65 0.40 0.55 0.9

Low Medium Low Low Low

2011/12 2011/12

Appendix C.6.1 Appendix C.6.2

Provision of partially controlled right-turns Change partial control to fully controlled right-turns Linked signals Signal visibility Linking of existing signals Replace a pedestal mount with mast arm mount signal Increase lens size to twelve inches Provide additional signal head Roundabouts Install roundabout rural Install roundabout urban Install roundabout all environments

10%

2011/12

Appendix C.6.2

70%

0.3

Low

2011/12

Appendix C.6.2

15% 35%

0.85 0.65

Medium Low

2008/09 2011/12

Austroads 2010a Appendix C.7

5% 20% 70% 55% 70% (all) 60% (pedestrians)

0.95 0.8 0.3 0.45 0.3 (all) 0.4 (pedestrians)

Low Medium High Medium High Low

2011/12 2011/12

Appendix C.7 Appendix C.7 Austroads 2010a

2008/09

Austroads 2010a Austroads 2010a

2010/11

Appendix C.18.7

Austroads 2012 7

Effectiveness of Road Safety Engineering Treatments

Treatment

Environment type All roundabouts lower risk for single lane compared with multilane (note this decision is usually not made based on safety, but rather on capacity) Install general channelisation Install splitter islands general Install splitter islands urban Install mountable median Install non-mountable median

Crash reduction factor 40%

Crash modification factor 0.6

Confidence

Year most recently assessed 2008/09

Reference location Austroads 2010a

Roundabout single versus multilane

Medium

Channelisation at intersections splitter and median islands

30% 30% 35% 15% 25%

0.70 0.70 0.65 0.85 0.75

Low Medium Medium Low Low

2011/12 2011/12 2011/12 2011/12 2011/12

Appendix C.8.1 Appendix C.8.2 Appendix C.8.2 Appendix C.8.3 Appendix C.8.3

Grade separation of intersections Install right-turn lane

X-intersection Y-intersection Install right-turn lane general Install right-turn lane signalised intersection Install right-turn lane unsignalised intersection Install right-turn lane rural unsignalised T-intersections Install right-turn lane rural unsignalised X-intersections Install right-turn lane urban: general Install right-turn lane urban unsignalised T-intersections Install right-turn lane urban signalised T-intersections

55% 20% 35% 30%

0.45 0.8 0.65 0.7

High Medium Medium Medium

2010/11 2010/11 2011/12 2011/12

Appendix C.9 Appendix C.9 Appendix C.10.1 Appendix C.10.2

35%

0.65

Medium

2011/12

Appendix C.10.3

40%

0.6

Low

2011/12

Appendix C.10.4

30%

0.7

Medium

2011/12

Appendix C.10.4

30%

0.7

Low

2011/12

Appendix C.10.5

35%

0.65

Low

2011/12

Appendix C.10.5

5%

0.95

Low

2011/12

Appendix C.10.5

Austroads 2012 8

Effectiveness of Road Safety Engineering Treatments

Treatment

Environment type Install right-turn lane urban unsignalised X-intersections Install right-turn lane urban signalised X-intersections Install right-turn lane painted Install right-turn lane physical

Crash reduction factor 30%

Crash modification factor 0.7

Confidence

Year most recently assessed 2011/12

Reference location Appendix C.10.5

Install right-turn lane (continued)

Low

10%

0.9

Low

2011/12

Appendix C.10.5

30% 35% 20% 35%

0.70 0.65 0.80 0.65

High Low Low Low

2011/12 2011/12 2011/12 2008/09

Appendix C.10.6 Appendix C.10.6 Appendix C.12 Austroads 2010a

Install left-turn lane All Staggered junctions With minor road traffic < 15% of main road With minor road traffic 1530% of main road With minor road traffic > 30% of main road Railway level crossing treatments From nothing to signage From signage to lights and bells From lights and bells to barriers From signage to barriers Improve sight distance Overtaking lanes Road resurfacing to improve skid resistance Sight distance improvements Improving/ correcting superelevation All All All All All

25%

0.75

Low

Austroads 2010a

35%

0.65

Low

Austroads 2010a

25% 50% 45% 70% 45%

0.75 0.5 0.55 0.30 0.55

Low Low Low Low Low

2008/09 2008/09 2008/09 2008/09 2008/09

Austroads 2010a Austroads 2010a Austroads 2010a Austroads 2010a Austroads 2010a

Road geometry and design treatments All All 25% 35% 0.75 0.65 Medium High 2010/11 2008/09 Appendix C.14 Austroads 2010a

Rural environments and intersections All

30%

0.7

Medium

2008/09

Austroads 2010a

SD* < 0.01 0.01 to < 0.02 0.02

CMFunction 1.00 (1.00 + 6(SD 0.01) (1.00 + 3(SD 0.02) 0.7

Low

2011/12

Appendix C.15

*SD = Superelevation deficiency

Sealing shoulders

All

30%

High

2008/09

Austroads 2010a

Austroads 2012 9

Effectiveness of Road Safety Engineering Treatments

Treatment

Environment type

Crash reduction factor

Crash modification factor 0.5 (all casualty) 0.30 (fatal) 0.6

Confidence

Year most recently assessed 2011/12 2008/09 2008/09

Reference location

Roadside treatments Install impact attenuators Install guardrail Access control (note that this is a relative risk that compares with the level of risk if there were no access points on an undivided road. It does not provide a CRF or CMF) All All All Open road 4-lane 50% (all casualty) 70% (fatal) 40% Medium High High Medium Austroads 2010a Austroads 2010a Appendix C.16

Relative Risk = 1+(0.02 x residential/km+0.10 x commercial/km+0.20 x minor junctions) x ({0.45 if median-solid or > 3 m}, 1 if no median) Relative Risk = 1+(0.01 x residential/km+0.05 x commercial/km+0.20 x minor junctions) x ({0.45 if median-solid or >3 m}, 1 if no median) Relative Risk = 1+(0.01 x residential/km+0.08 x commercial/km+0.05 x minor junctions) x ({0.45 if median-solid or > 3 m}, 1 if no median) Relative Risk = 1+(0.02 x residential/km+0.10 x commercial/km+0.20 x minor junctions) x ({0.45 if median-solid or > 3 m}, 1 if no median) 30% 25% 15% 40% 35% 20% 0.7 0.75 0.85 0.6 0.65 0.8

Open road 2-lane

Medium

2008/09

Austroads 2010a

Built up 4-lane

Medium

2008/09

Austroads 2010a

Built up 2-lane

Medium

2008/09

Austroads 2010a

Signage treatments (Advisory) Install bridge warning signs Install curve warning signs Guidance signs Speed advisory Vehicle activated signs Variable message signs Install fencing and barriers Improved lighting All Horizontal curves All All All All Low High Low Low Medium Medium 2008/09 2010/11 2008/09 2010/11 2008/09 2008/09 Austroads 2010a Appendix C.17.1 Austroads 2010a Appendix C.17.2 Austroads 2010a Austroads 2010a

Pedestrian treatments All All 20% (pedestrians) 60% (pedestrians at night) 50% (pedestrians) 35% (pedestrians) 85% (pedestrians) 20% (pedestrians) 0.8 (pedestrians) 0.4 (pedestrians at night) 0.5 (pedestrians) 0.65 (pedestrians) 0.15 (pedestrians) 0.8 (pedestrians) Medium Medium 2010/11 2010/11 Appendix C.18.1 Appendix C.18.2

Add pedestrian phase at signals Improve signal timing Install pedestrian overpass Raised (wombat) crossing

All All All All

Low Low Low Low

2010/11 2010/11 2010/11 2010/11

Appendix C.18.5 Appendix C.18.5 Appendix C.18.4 Appendix C.18.8

Austroads 2012 10

Effectiveness of Road Safety Engineering Treatments

Treatment

Environment type All All

Crash reduction factor 45% (pedestrians) 50% (pedestrians) 40% 20%(1) 20%(1) 30% 30%

Crash modification factor 0.55 (pedestrians) 0.5

Confidence

Year most recently assessed 2010/11 2008/09

Reference location Appendix C.18.6 Austroads 2010a

Refuge Rest on red

Medium Low

Speed and enforcement treatments Speed cameras Mobile overt Mobile covert urban Mobile covert rural Fixed overt urban Fixed overt rural Speed change in speed limit and change in speed Decreases in speed limit From no previous restriction to any speed limit 100 to 80 km/h 80 to 60 km/h 60 to 50 km/h All reductions in speed limit Increases in speed limit 100 to 110 km/h < 90 to > 90 km/h Any increase in speed limit Change in operating speed and effect on safety Where Speeda = speed after, and Speedb = speed before All environments Install red light camera Signalised intersection 25% increase 15% increase 15% increase 2 1 100% 1.25 1.15 1.15 2 Medium Medium Medium Medium 2008/09 20% 0.8 Medium 0.6 0.8 0.8 0.7 0.7 Medium Low Medium Medium Low 2008/09 2008/09 2008/09 2008/09 2008/09 2008/09 Austroads 2010a Austroads 2010a Austroads 2010a Austroads 2010a Austroads 2010a Austroads 2010a Austroads 2010a

15% 20% 20% 15%

0.85 0.8 0.8 0.85

Medium Medium Medium Medium 2008/09

Austroads 2010a Austroads 2010a Austroads 2010a Austroads 2010a Austroads 2010a Austroads 2010a Austroads 2010a Austroads 2010a Austroads 2010a

20% 5% 25%

0.80 0.95 0.75

Medium High Low

2008/09 2008/09 2011/12

Austroads 2010a Austroads 2010a Appendix C.19

Provide transverse Intersections rumble strips

Austroads 2012 11

Effectiveness of Road Safety Engineering Treatments

Treatment

Environment type

Crash reduction factor

Crash modification factor 0.85 0.55 0.45 0.70

Confidence

Year most recently assessed 2008/09 2008/09 2008/09 2008/09

Reference location

Traffic management treatments Medians Painted median all environments Built median urban Built median rural Median crossovers Convert bi- to uni-directional crossover Close bi-directional crossover Mid-block turning provisions No overtaking markings Parking ban Parking convert angle to parallel Street closure Street lighting All environments All All All Internal roads Peripheral roads New lighting all locations New lighting intersections New lighting mid-block New lighting rural intersection New lighting urban New lighting urban intersection New lighting motorway/ freeway interchange New lighting railway level crossing Upgrade lighting Traffic calming All environments 15%(2) 45% 55% 30% Low Medium Medium Medium Austroads 2010a Austroads 2010a Austroads 2010a Austroads 2010a

55%

0.45

Medium

2008/09

Austroads 2010a

35% 35% 20% 40% 40% 5% 35% (night) 50% (night) 40% (night) 30% (night) 30% (night) 30% (night) 50% (night)

0.65 0.65 0.8 0.6 0.6 0.95 0.65 (night) 0.5 (night) 0.6 (night) 0.7 (night) 0.7 (night) 0.7 (night) 0.5 (night)

Medium Low Low Low Medium Low Medium High High Medium Low Low High

2008/09 2008/09 2008/09 2008/09 2011/12 2011/12 2010/11 2010/11 2010/11 2010/11 2010/11 2010/11 2010/11

Austroads 2010a Austroads 2010a Austroads 2010a Austroads 2010a Appendix C.20 Appendix C.20 Appendix C.21.1 Appendix C.21.2 Appendix C.21.3 Appendix C.21.5 Appendix C.21.6 Appendix C.21.7 Appendix C.21.8

60% (night)

0.4 (night)

High

2010/11

Appendix C.21.9

35% (night) 20%

0.65 (night) 0.8

Medium Medium

2010/11 2008/09

Appendix C.21.10 Austroads 2010a

Austroads 2012 12

Effectiveness of Road Safety Engineering Treatments

Treatment

Environment type Increase crash risk with presence of road works Minor treatments at work zones (e.g. pavement marking) Major treatments (e.g. barriers, cones, markings, advanced warning signs)

Crash reduction factor 30% increase compared with when no works present 15% improvement over no treatment

Crash modification factor 1.3

Confidence

Year most recently assessed 2008/09

Reference location Austroads 2010a

Work zones

Low

0.85

Low

2008/09

Austroads 2010a

45% improvement over no treatment

0.55

Low

2008/09

Austroads 2010a

1 2

Covert speed camera evaluations are typically conducted on an area-wide basis, so cannot be compared to overt evaluations which are conducted at or near camera sites. Austroads 2010a provided a CRF range of 1520% for installation of a painted median. This has been rounded down to 15% to provide a single figure.

Notes: Crash Reduction/Crash Modification Factors provided in this table have been rounded to the nearest 5%. Treatments assessed in 2008/09 were evaluated as part of a previous Austroads project, Road safety engineering risk assessment: Part 6: Crash Reduction Factors (Austroads 2010a). This summarised research conducted between 2004 and 2009. Details on the studies assessed can be found in the appendices to Austroads (2010a).

Austroads 2012 13

Effectiveness of Road Safety Engineering Treatments

DISCUSSION AND RECOMMENDATIONS

Based on an extensive review of the relevant literature, crash reduction estimates have been developed for 57 treatment types. In total, 126 crash reduction values have been derived for these issues, as in some cases, information is available for different environment types (e.g. rural and urban), different crash types (e.g. all crashes and run-off-road) or different crash severities. Despite the extensive body of research on treatment effectiveness, there is still a lack of reliable information on many of the safety treatments assessed in this study. However, out of the 126 crash reduction values, 52% had a high or medium level of confidence. The additional work conducted through this study has increased the level of confidence in the treatment effectiveness compared to the results obtained in an earlier study on this topic (Austroads 2010a, where 40% of the results had medium or high levels of confidence). Further research is recommended for those crash reduction values that have a low level of confidence, or where there is no existing information. Such research will provide a higher level of confidence in the crash reduction estimates. As identified in this report, the most cost-effective approach is to conduct further reviews of literature, although it may be several years before adequate information exists. Alternatively, experimentation, or analysis from existing data sources (for instance, crash monitoring databases) will be required. A recommended approach to conducting evaluations using these approaches is provided in Austroads (2012). This guide also provides information to help determine the robustness of evaluations already undertaken, and the methodologies that they employ. Combinations of treatments are commonly used to address high risk sites. Very little information is available on the combined benefit from using more than one treatment. Previous Austroads research has identified the need to assess the benefits of groups of treatments (Austroads 2012). There is a need to prioritise the most commonly used treatment combinations, and then assess the effectiveness of these. In addition, ARRB in association with FHWA in the United States, initiated an international collaboration on the development of crash reduction values through the OECD Joint Transport Research Centre. Involvement in the OECD collaboration has provided access to an international panel of experts with extensive experience in the development of crash reduction values, and access to the information that they currently hold. In the longer term this will lead to a strengthened international collaboration on this topic, including the potential for joint projects to fill existing gaps in knowledge. Although confidence in some of the crash reductions provided is not high, these are still the best estimates available for each type of treatment. In the absence of additional information on crash reduction based on total casualty crashes it is recommended that these figures be considered for use by jurisdictions when calculating the expected benefits from treatments. It is important to continue collaboration with local and international research partners, as well as updating of information when relevant research is complete (such as current research on clear zones) to develop guidance on the effectiveness of further treatment types, as well as improving confidence in the values presented in this report.

Austroads 2012 14

Effectiveness of Road Safety Engineering Treatments

REFERENCES
Austroads 2009, Design, feasibility and application of an instrumented vehicle, AP-T145-09, Austroads, Sydney, NSW. Austroads 2010a, Road safety engineering risk assessment: Part 6: Crash Reduction Factors, AP-T151-10, Austroads, Sydney, NSW. Austroads 2010b, Road safety engineering risk assessment: part 1: relationships between crash risk and the standards of geometric design elements, AP-T146/10, Austroads, Sydney, NSW. Austroads 2012, An Introductory Guide for Evaluating Effectiveness of Road Safety Treatments, AP-R421-12, Austroads, Sydney, NSW. Elvik, R, Hye, A, Vaa, T & Srensen, M 2009, The handbook of road safety measures, 2 edn, Emerald Publishing, Bingley, UK. Hauer, E 2006, The frequency-severity indeterminacy [unreported crashes], Accident Analysis and Prevention, vol. 38, no. 1, pp.78-83. Hauer, E 2007, Kinds of safety evaluation study, Issues in road safety evaluation workshop, 28-29 March 2007, Romsey, Victoria, Monash University Accident Research Centre, Clayton, Vic. Hauer, E 2009, Speed and safety, Transportation Research Record, no. 2103, pp.10-7. Ogden, KW 1996, Safer roads: a guide to road safety engineering, Avebury Technical, Aldershot, UK.
nd

Austroads 2012 15

Effectiveness of Road Safety Engineering Treatments

APPENDIX A

PRIORITISING GAPS IN KNOWLEDGE

During the first stage of this project, a number of treatments or road safety features were identified where knowledge of effectiveness in reducing crashes was not adequate. This assessment was based on literature reviews (Austroads 2010a, 2010b) as well as stakeholder priorities (Austroads 2009). For each of these issues it was identified that there was a lack of sufficient research on the crash reduction effect, or that the research which did exist was not considered of suitable quality. A simple matrix was developed to help prioritise the identified gaps in knowledge, with each issue subjectively assessed by the project team against the following criteria:

Frequency (F) How often the treatment is likely to be used: When prioritising the identified gaps in knowledge, common treatments were given a higher priority (for example, signs). Cost (C) The typical cost of the treatment: It was considered that for high cost treatments, there is generally a greater need to be more certain about the safety benefit compared to a low cost treatment. In some cases low cost treatments tend to be very widely used, for example signs. Although the individual cost of this treatment is relatively low, the sheer number of signs installed and maintained means that there is a high total cost from the use of this treatment. Uncertainty (U) The level of uncertainty associated with the known crash reduction effectiveness of the treatment: Hauer (2007) suggests that there are costs associated with making incorrect assumptions about the treatment effectiveness. The likelihood of this error can be calculated given the standard error and expected crash reduction.

For each of these factors a one to five scale was adopted, with one indicating a low value, and five indicating a high value. A qualitative approach was taken in the prioritisation of gaps in knowledge, based on subjective assessment of each of the above factors. It was noted that a quantitative approach could have been adopted (involving development of a benefit-cost ratio for each of the treatment types, allowing direct comparison between them). However, it was decided that such a quantitative approach would require a number of assumptions, and that therefore the additional analysis required was unlikely to produce a more robust result than the qualitative approach. The ratings for each were then multiplied to give a total score 3. Table A 1 provides the results of this assessment.
Table A 1:
Treatment Channelisation at intersections splitter and median islands Clear zones Delineation continuity lines, marking in wet conditions Grade separation of intersections Frequency (F) 2

Gaps in knowledge priority matrix


Cost (C) 3 Uncertainty (U) 4 Comments Splitter islands mainly from one NZ study. Median islands based on one US study. Very high priority, but being addressed through ST1427. No information available. Based on one study only (Elvik et al. 2009). No Australasian figure available. FxCxU 24

5 4 2

5 2 5

5 5 5

125 40 50

An alternative prioritisation method was also applied whereby frequency and cost are added together before being multiplied by uncertainty. However, the outcome from this analysis was similar to a straight multiplication of each element hence the method was not included in this report.

Austroads 2012 16

Effectiveness of Road Safety Engineering Treatments

Treatment Guide posts

Frequency (F) 5

Cost (C) 2

Uncertainty (U) 3

Comments Five studies, with one from Australia and one from NZ. Gap relates to effectiveness on curves versus straights. Two US studies available, but robustness uncertain. One Australian figure available, but cannot trace origin of this, and so therefore cannot use. Based on two US studies. Three studies, one from Australia. Delete from list adequate information on this. Many studies on this topic, and the overall CRF is quite reliable. Gaps in knowledge regarding different types of intersection. Only one US study available. No information on this topic. Gap in knowledge actually relates to centreline markings. Only one Canadian study available as part of original review. There is now more recent literature on this. Four US studies available, no Australasian study. Robustness of US studies uncertain. Adequate information regarding overtaking lane provision. Gaps appear when looking at specific lane types. Seven studies assessed, but many with low robustness. No studies from Australasia. Three estimates provided (two US, Elvik et al.). Estimates vary substantially leading to low confidence. No estimates from Australasia. Based on six studies, but none from Australasia. Also, no information on use at curves. Gap in knowledge relates to provision at curves. Several marking types; no Australasian studies. Four studies, two from Australia.

FxCxU 30

Impact attenuators

27

Intersection extend right-turn lane Intersection left-turn provision Intersection red light camera Intersection right-turn lane

2 3 3 3

3 3 3 3

4 3 1 2

24 27 9 18

Intersection signal visibility Lane width tunnels Line marking profile edge lines

3 2 2

2 4 2

4 5 4

24 40 16

No overtaking markings

12

Overtaking lanes

36

Parking ban

Parking convert angle to parallel

Pavement markings centreline

30

Pavement markings edge line Pavement markings words and symbols Pavement markings painted speed limits Pedestrian crossings Pedestrian treatment rest on red Railway crossing treatments

5 1 1 5 1 2

2 1 1 3 1 4

3 4 3 5 4 3

30 4 3 75

One Australian study. Three studies Elvik et al. and two from Australia. However, Australian studies do not provide source. Gaps in speed advisory signs, curve warning, chevrons, animal warning.

4 24

Signs advisory

60

Austroads 2012 17

Effectiveness of Road Safety Engineering Treatments

Treatment Signs regulatory

Frequency (F) 5

Cost (C) 3

Uncertainty (U) 4

Comments Gaps relating to installation of stop and give-way signs. Although there are a number of studies, each gives vastly different results. Currently based on Elvik et al. Australian data has been identified. Ogden (1996) provides an estimate, but basis of this uncertain. Gap in knowledge only appears when looking at very specific environments. Several studies found, but none thought to be reliable. There are several studies on this topic, two significant ones from Australia. Only really low confidence because CRFs are broken down by different types of signal installations. One US study plus the Austroads guide which makes an estimate. Currently being addressed in clear zone research. Seven studies on increased risk at work zones, and results are consistent. Gap in knowledge lies in severity outcomes from different treatments.

FxCxU 60

Staggered junctions Street closure Street lighting Superelevation Traffic signals

2 1 4 4 5

4 3 4 3 3

2 5 2 5 2

16 15 32 60 30

U-turn bans Vegetation Work zones

2 4 3

1 4 3

5 5 3

10 80 27

Austroads 2012 18

Effectiveness of Road Safety Engineering Treatments

APPENDIX B

REVISIONS TO TREATMENT EFFECTIVENESS FIGURES

Table B 1 provides a comparison summary of treatment effectiveness values that have been updated since the completion of Road Safety Engineering Risk Assessment: Part 6: Crash Reduction Factors (Austroads 2010a). Austroads 2010a followed the same methodology as the current project, and involved literature reviews of 47 treatment types for which 104 treatment effectiveness values were determined. For these, 40% of the results had medium or high levels of confidence attributed. (Comparatively, for the current project, 126 crash reduction values were derived for 57 treatment types, for which 52% had a high or medium level of confidence.)
Table B 1: Updated treatment effectiveness values
Revised descriptions or reductions identified in current project Treatment type Channelisation at intersections splitter and median islands Environment type Install general channelisation Install splitter islands general Install splitter islands urban Install mountable median Install non-mountable median % Reduction 30% 30% 35% 15% 25% Confidence Low Medium Medium Low Low

Original Crash Reduction Factors identified in Road safety engineering risk assessment: Part 6: Crash Reduction Factors (Austroads 2010a) Issue Channelisation at intersections splitter and median islands Environment type Splitter island all environments Splitter island rural Splitter island urban Splitter island T-intersection Splitter island X-intersection Median island mountable Median island non-mountable Grade separation of intersections Guide posts Impact attenuators All environments, grade separation of intersection Rural All % Reduction 40% 35% 40% 45% 40% 15% 25% 50% (all severities) Confidence Low Low Low Low Low Low Low Low Grade separation of intersections Install guide posts

X-intersection Y-intersection All All All

55% 20% 5% 50% (all casualty) 70% (fatal)

High Medium Low Medium High

28% (night) 60%

Low Low

Install impact attenuators

Austroads 2012 19

Effectiveness of Road Safety Engineering Treatments

Original Crash Reduction Factors identified in Road safety engineering risk assessment: Part 6: Crash Reduction Factors (Austroads 2010a) Issue Intersection right-turn lane Environment type All environments At signalised intersections At unsignalised intersections Urban Rural Painted Protected % Reduction 35% 35% 35% 30% 35% 30% 35% Confidence Medium Low Low Low Low Low Low Treatment type Install right-turn lane

Revised descriptions or reductions identified in current project Environment type Install right-turn lane general Install right-turn lane signalised intersection Install right-turn lane unsignalised intersection Install right-turn lane rural unsignalised T-intersections Install right-turn lane rural unsignalised X-intersections Install right-turn lane urban general Install right-turn lane urban unsignalised T-intersections Install right-turn lane urban signalised T-intersections Install right-turn lane urban unsignalised X-intersections Install right-turn lane urban signalised X-intersections Install right-turn lane painted Install right-turn lane physical % Reduction 35% 30% 35% 40% 30% 30% 35% 5% 30% 10% 30% 35% Unknown 20% Confidence Medium Medium Medium Low Medium Low Low Low Low Low High Low n/a Low

Intersection extend right-turn lane Intersection left-turn provision

All environments All environments

15% 30%

Low Low

Intersection extend right-turn lane Install left-turn lane

Unknown All

Austroads 2012 20

Effectiveness of Road Safety Engineering Treatments

Original Crash Reduction Factors identified in Road safety engineering risk assessment: Part 6: Crash Reduction Factors (Austroads 2010a) Issue Intersection signal visibility Environment type All environments % Reduction 25% Confidence Low Treatment type Signal visibility

Revised descriptions or reductions identified in current project Environment type Replace a pedestal mount with mast arm mount signal Increase lens size to twelve inches Provide additional signal head % Reduction 35% 5% 20% 20% (all) 40% (run-off-road) 15% (all) 30% (head-on) 25% 25% Confidence Low Low Medium Medium Low Medium Low Low Medium

Line markings profile edge lines

Shoulder Centreline

23% 15% Unknown 23% 5% 54% 30% 20% Unknown 0%

Medium Low n.a. Medium Low Low Low Low n.a. Low

Profile line marking

Provide profile edge lines Provide profile centreline

Line markings profile edge lines Overtaking lanes

Transverse All environments Tack-on lane New alignment and passing lane

Provide transverse rumble strips Overtaking lanes

Intersections All

Pavement markings

Centreline Edge line Words and symbols Painted speed limits

Pavement markings

Provide edge line Provide centreline Provide combined edge and centreline Provide painted speed limits

10% 20% 30% 0%

Low Low Low Low

Austroads 2012 21

Effectiveness of Road Safety Engineering Treatments

Original Crash Reduction Factors identified in Road safety engineering risk assessment: Part 6: Crash Reduction Factors (Austroads 2010a) Issue Pedestrian crossings Environment type No reliable reductions determined for this issue % Reduction Unknown Confidence n.a. Treatment type

Revised descriptions or reductions identified in current project Environment type Install fencing and barriers Improved lighting Add pedestrian phase at signals Improve signal timing % Reduction 20% (pedestrians) 60% (pedestrians at night) 50% (pedestrians) 35% (pedestrians) 85% (pedestrians) 20% (pedestrians) 45% (pedestrians) 50% (pedestrians) 60% (pedestrians) 15% 30% 60% 25% 60% Confidence Medium Medium Low Low Low Low Medium Low Low Low Medium Low Low Medium

Pedestrian treatments

Pedestrian treatment rest on red

All

50%

Low

Install pedestrian overpass Raised (wombat) crossing Refuge Rest on red Install roundabout all environments

Signs regulatory (intersection)

Install stop sign at T-intersection Install stop sign at X-intersection Install give-way sign all intersections

20% 35% 15% Unknown 25% 30%

Medium Medium Low n.a. Low Low

Signs regulatory

Install stop sign at T-intersection Install stop sign at X-intersection Install four-way stop sign at X-intersection Install give-way sign all intersections Install right-turn ban, or U-turn and right-turn ban

U-turn bans Signs regulatory (midblock) Signs advisory

No reliable reductions determined for this issue All environments Chevron warning signs all environments

Treatment type omitted, as not considered useful Delineation treatments install chevrons Horizontal curves 25% Low

Austroads 2012 22

Effectiveness of Road Safety Engineering Treatments

Original Crash Reduction Factors identified in Road safety engineering risk assessment: Part 6: Crash Reduction Factors (Austroads 2010a) Issue Signs advisory Environment type Advisory speed signs all environments Curve warning signs all environments Street closure Street lighting No reliable reductions determined for this issue Install lighting midblock Install lighting rural Install lighting rural intersection Install lighting urban Install lighting urban intersection % Reduction 25% 25% unknown 40% (night) 30% (night) 40% (night) 30% (night) 20% (night) Confidence Low Low n.a. Medium Low Medium Low Low Street closure Street lighting Treatment type Advisory signage treatments

Revised descriptions or reductions identified in current project Environment type Install curve warning signs Speed advisory Internal roads Peripheral roads New lighting all locations New lighting intersections New lighting mid-block New lighting rural intersection New lighting urban New lighting urban intersection New lighting motorway/freeway interchange New lighting railway level crossing Upgrade lighting Improving/correcting superelevation All % Reduction 25% 40% 40% 5% 35% (night) 50% (night) 40% (night) 30% (night) 30% (night) 30% (night) 50% (night) 60% (night) 35% (night) 10% Confidence High Low Medium Low Medium High High Medium Low Low High High Medium Low

Superelevation

No reliable reductions determined for this issue

Unknown

n.a.

Austroads 2012 23

Effectiveness of Road Safety Engineering Treatments

Original Crash Reduction Factors identified in Road safety engineering risk assessment: Part 6: Crash Reduction Factors (Austroads 2010a) Issue Traffic signals Environment type New signals, no turn arrows metro New signals, with turn arrows metro New signals, no turn arrows regional New signals, with turn arrows regional % Reduction 45% 40% 75% Confidence Low Low Low Treatment type Traffic signals

Revised descriptions or reductions identified in current project Environment type Install traffic signals Provision of fully controlled rightturns % Reduction 30% 35% (all casualty) 60% (right through) 45% (adjacent direction) Confidence Low Medium Low Low Low Low

35%

Low

Provision of partially controlled right-turns Change partial control to fully controlled right-turns

10% 70%

Vegetation Weather

No reliable reductions determined for this issue Dry weather, reduction from crash risk in wet weather

Unknown 20%

n.a. Low

Treatment type omitted (topic is being covered as part of other current research) Treatment type omitted (topic is being covered as part of other current research)

Austroads 2012 24

Effectiveness of Road Safety Engineering Treatments

APPENDIX C

SUMMARY OF RESEARCH FOR PRIORITY TREATMENTS

This appendix presents details of the research conducted in 2010/11 and 2011/12. Each appendix includes a tabulated summary of treatment effectiveness (generally reductions) found in the research. Where available, effectiveness for different levels of severity or crash types were included. The main focus for the review was on recent research although in some cases older research is also included. There was also a focus on research conducted in Australia and New Zealand. However, other key references have also been assessed and included where relevant. Only research that contained information about the crash reduction benefit from treatments was included, which greatly reduced the number of relevant publications. Note that throughout this appendix many of the references assessed did not contribute any information to the final crash reduction factor. The references are provided for completeness, but may also provide a useful source of additional information on each treatment type. A discussion is included for each which presents the rationale for the treatment effectiveness values presented in Table 3.1. The review identified some evidence for a number of additional treatments (than those presented in Table 3.1) for which there was insufficient evidence to determine the extent of treatment effectiveness (e.g. provision of edge lines at curves, addition of yellow reflective tape to signal heads, and channelisation in rural environments). In such cases, the relevant appendix generally includes a discussion indicating that safety benefits are likely to be associated with such treatments, although there was insufficient evidence to gauge treatment effectiveness during this project.

Austroads 2012 25

Effectiveness of Road Safety Engineering Treatments

C.1

Guide Posts
Study Year 1978 1978 Country Australia USA Environment/Treatment Open General (all sites) Reduction Corner cube showed 60% reduction in night, only 21% reduction at control No treatment: 3.2943 (standard error 0.511) With centreline: 2.3473 (standard error 0.114) With centreline & post: 1.3285 (standard error 0.080) No treatment: 3.7740 (standard error 0) With centreline: 2.2375 (standard error 0.169) With centreline & post: 1.1323 (standard error 0.071) Centreline and edge line: 2.4925 (standard error 0.261) Centreline and edge line and post: 1.9306 (standard error 0) Corner cube 15% decrease at night, but 21% decrease at control sites Post delineators 20% reduction 3267% reduction on curves, 1518% on whole route Open Corner cube delineators (17 before, 6 after) 65% reduction Recommended 30% reduction in night-time Reported post delineators at curves 23% reduction in all; 30% at night Reported post delineators at tangent 1628% reduction in all; 30% reduction in night Reported flexible delineator post 40% reduction in all 27% reduction, but only at one site (not significant). Crash costs reduced by 85%, and were significant. Two-lane roads Undivided roads Divided roads All accidents All accidents Combined treatment of edge line, centreline and delineator posts 30% reduction for delineators on curves 67% reduction 30% reduction Injury accidents 7% reduction (95% CI -22; +12) PDO 3% reduction (95% CI -27; +28) Injury accident 45% reduction (95% CI -56; -32)

Summary of research
Vincent Bali et al.

Bali et al.

1978

USA

Straight sites

Bali et al.

1978

USA

Curves

Sanderson & Fildes Creasey & Agent LTSA Cairney Agent, Stamatiadis & Jones Agent, Stamatiadis & Jones Agent, Stamatiadis & Jones Agent, Stamatiadis & Jones Corben et al.

1984 1985 1992 1993 1996 1996 1996 1996 1997

Australia (Victoria) USA NZ Australia (Victoria) USA USA USA USA Australia

Open

Douglas Douglas Douglas Elvik et al. Elvik et al. Elvik et al.

2000 2000 2000 2009 2009 2009

USA UK UK International International International

Austroads 2012 26

Effectiveness of Road Safety Engineering Treatments

Study AASHTO

Year 2010

Country USA

Environment/Treatment Installing post-mounted delineators on rural two-lane undivided roads Rural two-lane road

Reduction Injury accidents have a CMF of 1.04 (caveat observed variability suggests this treatment could result in an increase, decrease or no change) Combined treatment Edge lines, centrelines and guide posts Injury crashes have a CMF of 0.55 (standard error 0.1)

AASHTO

2010

USA

Results from older studies have not been used given that new guide post designs now exist and are likely to provide a different benefit. The study by Corben et al. is from Australia, but did not provide a statistically significant result, although it does provide a useful direction of effect (i.e. a possible positive benefit). The Vincent review is based on older research and was also excluded. Elvik et al. and the AASHTO provide more recent results. Elvik et al. suggests a 7% CRF for the treatment while the AASHTO suggests there is a 4% increase in crashes (although it is observed that due to variability, this treatment could deliver an increase, decrease, or no change in crashes). It appears from more recent research that the benefits from using guide posts on their own might not be substantial (and indeed, the US research indicates a possible increase in crashes, most likely due to an increase in speed which has been identified in research). However, additional assessment by Elvik et al. suggests that when used in combination with other delineation treatments (specifically an edge and centreline) that the benefit is likely to be substantial (a 45% reduction). Agent, Stamatiadis and Jones indicate that there is no great difference in the benefit of guide posts at curves versus straights, although the New Zealand research from LTSA indicates that benefits are substantially greater at curves when compared to whole routes. Given the variability of the results from the HSM, the Elvik et al. figure of a 7% reduction is currently recommended (rounded to 5%). The crash reductions provided by Elvik et al. are specified as all accidents. Some of the research indicates that benefits at night are greater than changes in at all times of day for this treatment type, however there is insufficient information to provide a figure. References assessed
AASHTO 2010, Highway safety manual, 1st edn, American Association of State Highway and Transportation Officials, Washington, DC, USA. Agent, KR, Stamatiadis, N & Jones, S 1996, Development of accident reduction factors, report KTC-96-13, Kentucky Transportation Centre, University of Kentucky, Lexington, KY, USA. Bali, S, Potts, R, Fee, JA, Taylor, JI & Glennon, J 1978, Cost effectiveness and safety of alternative roadway delineation treatments for rural two-lane highways: vol II: final report, FHWA/RD-78-51, Federal Highway Administration, Washington, DC, USA. Cairney, P 1993, Current issues in delineation, Road and Transport Research, vol.2, no.2, pp.28-39. Corben, B, Deery, H, Mullan, N & Dyte, D 1997, The general effectiveness of countermeasures for crashes into fixed roadside objects, report 111, Monash University Accident Research Centre, Clayton, Vic.

Austroads 2012 27

Effectiveness of Road Safety Engineering Treatments

Creasey, T & Agent, KR 1985, Development of accident reduction factors, research report UKTRP-85-6, Kentucky Transportation Research Program, University of Kentucky, Lexington, KY, USA. Douglas, J 2000, Making the delineation message clear, Roadmarking Industry Association of Australia and New Zealand Roadmarkers Federation joint conference, 6th, 2000, Canberra, Australian Capital Territory, Roadmarking Industry Association of Australia, Rosebud, Vic, 16 pp. Elvik, R, Hye, A, Vaa, T & Srensen, M 2009, The handbook of road safety measures, 2 Publishing Group, Bingley, UK.
nd

edn, Emerald

Land Transport Safety Authority 1992, Guidelines for rural road marking and delineation, RTS 5, New Zealand Transport Agency, Wellington, NZ, viewed 14 July 2011, <http://www.nzta.govt.nz/resources/road-traffic-standards/docs/rts-05.pdf>. Montella, A 2005, Safety reviews of existing roads: a quantitative safety assessment methodology, Transportation Research Record, no. 1922, Transportation Research Board, Washington, pp. 65-72. Sanderson, JT & Fildes, B 1984, Run-off-the-road accidents in rural areas, report TS84/6, Traffic and Safety Department, Royal Automobile Club of Victoria (RACV), Melbourne, Vic. Vincent, EN 1978, A trial installation of corner cube delineators: Calder Highway, Gisborne to Woodend, Australian Road Research, vol. 8, no. 3, pp. 38-40.

Austroads 2012 28

Effectiveness of Road Safety Engineering Treatments

C.2

Chevron Alignment Markers


Study Year 1996 Country USA Environment/Treatment Not specified Reduction Warning signs chevron 30% average reduction all crashes (lit review three papers) 55% average reduction all crashes (state survey two states) 46.3% (95% CI -171.5; 78.8). Not significant all severities, non-intersection CMF 4% all severities, non-intersection, head-on, run-off-road, side swipe CMF 6% fatal/serious injury/minor injury, non-intersection CMF 16% all severities, night, non-intersection CMF 25% all severities, night, non-intersection, head-on, run-off-road, side swipe CMF 22%

Summary of research
Agent, Stamatiadis and Jones

Montella Srinivasan et al.

2009 2009

Italy USA

Divided highways, 4 lanes total Install chevrons on horizontal curves rural

AASHTO

2010

USA

Installing chevron signs on horizontal curves along urban and suburban arterials Rural Rural but with other supporting treatments Any environment, and when used with other supporting treatments

appears to reduce crashes of all types. However, the magnitude of the crash effect is not certain at this time. (p 1366) 35% reduction (95% CI -1.28; 70.87) 46% reduction (95% CI 21.95; 70.4) 50% reduction (95% CI 34.88; 64.46)

UK-MoRSE website

2010

UK

A CRF of 25% was selected based on the limited evidence available. This includes a 30% reduction from Agent, Stamatiadis and Jones, a 4% reduction from Srinivasan et al. and a 35% reduction from the UK MoRSE database. There were also non-significant findings from Montella and AASHTO that this treatment has a positive benefit. There is a low level of confidence in this figure. References assessed
AASHTO 2010, Highway safety manual, 1st edn, American Association of State Highway and Transportation Officials, Washington, DC, USA. Agent, KR, Stamatiadis, N & Jones, S 1996, Development of accident reduction factors, research report KTC-96-13, Kentucky Transportation Centre, University of Kentucky, Lexington, KY, USA. Andreassen, DC 1989, Strategies for safety problems, research report ARR 163, Australian Road Research Board, Vermont South, Vic. Andrew OBrien and Associates 2000, Managing traffic flow on urban freeways: appendix A: literature review, Andrew OBrien and Associates, Melbourne, Vic.

Austroads 2012 29

Effectiveness of Road Safety Engineering Treatments

Austroads 2009, Guide to road safety: part 8: treatment of crash locations, AGRS08/09, Austroads, Sydney, NSW. Bureau of Transport Economics (BTE) 2001, The black spot program 1996-2002: An evaluation of the first three years, report no. 104, Bureau of Transport Economics (BTE), Canberra, Australia. Creasey, T & Agent, KR 1985, Development of accident reduction factors, research report UKTRP-85-6, Kentucky Transportation Research Program, University of Kentucky, Lexington, KY, USA. Donald, D 1997, Be warned! A review of curve warning signs and curve advisory speeds, research report 304, Vermont South, Vic. Elvik, R, Hye, A, Vaa, T & Srensen, M 2009, The handbook of road safety measures, 2 Publishing Group, Bingley, UK.
nd

edn, Emerald

Ewing, R 1999, Traffic calming: state of the practice, report FHWA-RD-99-135, Federal Highway Administration, Washington, DC, USA. Kneebone, DC 1964, Advisory speed signs and their effect on traffic, Australian Road Research Board conference, 2nd, 1964, Melbourne, Victoria, Australian Road Research Board, Vermont South, Vic, vol. 2, no. 1, pp. 524-41. Kulmala, R 1994, Measuring the safety effect of road measures at junctions, Accident Analysis and Prevention, vol. 26, no. 6, pp. 781-94. Lamm, R, Zumkeller, K & Beck, A 2001, Traffic safety: the relative effectiveness of a variety of road markings and traffic control devices, Road Safety on Three Continents, 2000, Pretoria, South Africa, VTI Konferens 15A, Swedish National Road and Transport Research Institute, Linkoeping, Sweden, pp. 120-32. Montella, A 2009, Safety evaluation of curve delineation improvements: empirical Bayes observational before-and-after study, Transportation Research Record, no. 2103, Transportation Research Board, Washington, pp. 6979. Moses, P 1987, Combating the road toll, National Local Government Engineering conference, 4th, 1987, Perth, Western Australia, Institution of Engineers Australia, Canberra, ACT, pp.70-4. Srinivasan, R, Baek, J, Carter, D, Persaud, B, Lyon, C, Eccles, K, Gross, F, & Lefler, N 2009, Safety evaluation of improved curve delineation, report FHWA-HRT-09-045, Federal Highway Administration, Washington, DC, viewed 13 December 2010, <http://www.fhwa.dot.gov/publications/research/safety/09046/>. UK-MoRSE 2010, UK-MoRSE, Greensafe Foundation, Birmingham, UK, viewed 15 December 2010, <www.uk-morse.com>. Winnett, MA & Wheeler, AH 2002, Vehicle-activated signs: a large scale evaluation, report 548, TRL, Crowthorne, UK.

Austroads 2012 30

Effectiveness of Road Safety Engineering Treatments

C.3
C.3.1

Pavement Markings
Provision of Edge Lines
Study Year 1981 1984 1985 1986 Country UK UK USA Australia Rural roads Not specified Rural highways Environment/Treatment Reduction 13% to 32% reduction all crashes 37% to 42% reduction night crashes Inconclusive 15% reduction total crashes Wide edge lining: 8% reduction all crashes 34% reduction out-of-control single vehicle 15% reduction straight roads 45% reduction curves 0% no effect for run-off-road crashes Only effective at sites that experience loss of control type crashes Curves 25% reduction night time crashes 5% reduction day time crashes for failing to make the bend crashes through addition of RRPMs and edge lining on lower speed curves 10% reduction delineation (does not specify type) 20% reduction (average) all crashes 2% reduction all crashes 25% reduction run-off-road Not specified 2021% reduction average for delineation Edge lines: 8% reduction USA nationwide 16.5% reduction Kansas 14.5% reduction Kansas 19% reduction Ohio 21% reduction Illinois 16% reduction Idaho 38% reduction Utah 60% reduction Arizona 3% reduction Michigan 18% reduction East Sussex 30% reduction South Yorkshire 26% reduction Cornwall 12% reduction Northamptonshire 22% reduction Hertfordshire 27% reduction Lorraine 20% reduction Hesse 25% reduction Lower Saxony 36% reduction Centreline and edge line (where previously none) 8% reduction add edge line where already centreline

Summary of research

Jackson in Ogden (1996) Willis, Scott and Barnes in Ogden (1996) Creasey & Agent Moses in in Ogden (1996) Nairn in Ogden (1996) Cottrell County Surveyors Society in Ogden (1996) Moses

1987 1987 1989 1990

Australia USA UK Australia

Straight roads Not specified

VicRoads Miller in in Ogden (1996) Tignor in in Ogden (1996) Miller

1990 1992 1993 1993

Australia USA USA USA and others

Curves

Austroads 2012 31

Effectiveness of Road Safety Engineering Treatments

Study Moses

Year 1994

Country Australia (Western Australia) USA

Environment/Treatment Not specified

Reduction 25% reduction fatal out-of-control crashes with provision of 150 mm edge lines 20% reduction state survey 15% reduction lit review 15% reduction all crashes recommended 30% reduction off-road crashes recommended 1.4% reduction edge line (not statistically significant) 2 crashes per 106 VKT for varying curve radius Varies between 0.75 crashes per 106 VKT (25 gon/km change in curve rate) to 2.25 crashes per 106 VKT (450 gon/km change in curve rate) Varies between 14 crashes per 106 VKT (5 m pavement) to 2 crashes per 106 VKT (7 m pavement) 10% to 15% reduction for run-off-road crashes at high crash sites Edge lines onto rural roads East Sussex treatment vs. control 18% reduction in all crashes no change in daylight crashes 43% reduction in night-time crashes Edge lines onto rural roads South Yorkshire treatment vs. control 30% reduction in all crashes No change in daylight crashes 68% reduction in night-time crashes (there was 38% reduction at treatment sites and 29% increase at control sites) Normal edge line [note no indication if straight or curve] Injury crashes reduced by 3% (95% CI -7; +1) Wide edge line (20 cm instead of 10 cm) [note no indication if straight or curve] Injury crashes increased by 5% (95% CI -4; +14) On rural 2 lane road edge lines [note no indication if straight or curve] Injury crash CMF is 0.97 (standard error 0.04) On rural 2 lane road wide edge lines (8 inches or 20.3 cm) [note no indication if straight or curve] Injury crashes CMF of 1.05 (standard error 0.08)

Agent, Stamatiadis & Jones

1996

Not specified

Corben et al. Lamm, Zumkeller & Beck

1997 2001

Australia South Africa

Not specified Curves

Lamm, Zumkeller & Beck Council et al. Roadway Safety Markings Association

2001

South Africa

Rural

2002 2007

USA UK

Two-lane roads Rural roads

Roadway Safety Markings Association

2007

UK

Rural roads

Elvik et al.

2009

Intl

Elvik et al.

2009

Intl

AASHTO

2010

US

AASHTO

2010

US

Austroads 2012 32

Effectiveness of Road Safety Engineering Treatments

A crash reduction of 10% for all casualty crashes was based on the results from Creasy and Agent (15%), County Surveyors Society (only effective for loss of control type crashes), Willis et al. (0%), Agent et al. (15%), Jackson (average of 22.5%), Moses (8%), Miller (20%), Tignor (2%), Corben et al. (1.4%), Elvik et al. (3%), RSMA (18%, 30%). Given the range of results, and indications for more recent research that there is minimal effect on safety, there is low confidence in this figure. There are indications that the installation of edge lines has a minimal effect on safety, with the results from the Roadway Safety Markings Association showing no benefit during daytime. Benefits for night-time crashes are therefore likely to be higher than the 10% average provided above. There is some information on the crash reduction benefit of edge lines for run-off-road crashes. The mean reduction was 15%, based on the average of Cottrell (0%), Council et al. (10% to 15%), Agent, Stamatiadis and Jones (30%) and Tignor (25%). Low confidence is placed in this CRF as there are few studies and the range is wide. The crash reduction for wide edge lines is unable to be determined as there are only two studies which have different outcomes: Moses (in Douglas 2000) study had an 8% reduction, whereas Elvik et al. had a 5% increase. There is some indication of a greater benefit of edge lines at curves, but the information is not reliable enough to provide a figure. C.3.2
Bali et al.

Provision of Centrelines
Study Year 1978 Country USA Environment/Treatment All Reduction Note this is accident rate (per million vehicle miles) and not just injury crashes General sites no treatment (mean: 3.29, standard error: 0.511) vs. painted centreline (mean: 2.35, standard error: 0.114) Not significant Note this is accident rate (per million vehicle miles) and not just injury crashes Straights no treatment (mean: 3.77, standard error not reported) vs. painted centreline (mean: 2.24, standard error: 0.152). Significant at 0.05 level Note this is accident rate (per million vehicle miles) and not just injury crashes Curves centre and edge line (mean: 2.49, standard error: 0.261) vs. centre and edge line and guide posts (mean: 1.96, standard error: not reported) Significant at 0.05 level 30% reduction centreline 40% reduction no passing striping 10% reduction delineation overall 65% reduction barrier line

Bali et al.

1978

USA

Straight roads

Bali et al.

1978

USA

Curves

Creasey & Agent VicRoads

1985 1990

USA Australia

Not specified Curves

Austroads 2012 33

Effectiveness of Road Safety Engineering Treatments

Study Miller

Year 1993

Country USA and others

Environment/Treatment Not specified

Reduction 2021% reduction average for delineation Centrelines: 29% reduction for USA 10% reduction for Bavaria 36% reduction centreline and edge line (where previously none) 36% reduction state survey 24% reduction lit review 35% reduction recommended Varies between 14 crashes per 106 VKT (5 m pavement) to 2 crashes per 106 VKT (7 m pavement) Essentially 2 crashes per 106 VKT for varying curve radius) Varies between 0.75 crashes per 106 VKT (25 gon/km change in curve rate) to 2.25 crashes per 106 VKT (450 gon/km change in curve rate) Injury crashes reduced by 1% (95% CI -8; +6) Injury crashes have a CMF of 0.99 (standard error 0.06)

Agent, Stamatiadis & Jones Lamm, Zumkeller & Beck Lamm, Zumkeller & Beck

1996

USA

Not specified

2001

South Africa

Rural, straight

2001

South Africa

Rural, curves

Elvik et al.

2009

International

Centreline (note no indication if straight or curve) On rural 2 lane road centrelines (note no indication if straight or curve)

AASHTO

2010

USA

There is a mean crash reduction of 20% based on the average reduction of Miller (29%, 10%), Creasy and Agent (30%), Agent et al. (35%) and Elvik et al. (1%). There is a low level of confidence in this estimate due to the discrepancy between the figure given by Elvik et al. and the other three studies. C.3.3
Bali et al.

Provision of Centreline and Edge Line


Study Year 1978 Country USA Environment/Treatment Curves Centre and edge line Centre and edge line and guide posts Reduction Note this is accident rate (per million vehicle miles) and not just injury crashes Mean: 2.49, standard error 0.261 Mean: 1.96, standard error not reported Significant at 0.05 level 36% reduction centreline and edge line (where previously none) Injury accident 45% reduction (95% CI -56 -32)

Miller Elvik et al.

1993 2009

USA and others International

Not specified Combined treatment of edge line, centreline and delineator posts Combined treatment of edge line and centreline On rural 2 lane road edge and centrelines On rural 2 lane road edge line centrelines and guide posts

Elvik et al. AASHTO AASHTO

2009 2010 2010

International US US

Injury accident 24% reduction (95% CI -35 -11) Injury crashes have a CMF of 0.76 (standard error 0.1) injury crashes have a CMF of 0.55 (standard error 0.1)

Austroads 2012 34

Effectiveness of Road Safety Engineering Treatments

There is a mean crash reduction of 30%, based on Miller (36%) and Elvik et al. (24%). Note that the AASHTO repeats the information from Elvik et al. so it has not been included here. Low confidence is placed in this estimate as it uses only two studies. Elvik et al. notes that there is an even higher benefit when guide posts are also included. References assessed
AASHTO 2010, Highway safety manual, 1 edn, American Association of State Highway and Transportation Officials, Washington, DC, USA. Agent, KR, Stamatiadis, N & Jones, S 1996, Development of accident reduction factors, research report KTC-96-13, Kentucky Transportation Centre, University of Kentucky, Lexington, KY, USA. Austroads 2001, Economic evaluation of road investment proposals: improved prediction models for road crash savings, APR184/01, Austroads, Sydney, NSW. Austroads 2009, Guide to road safety: part 8: treatment of crash locations, AGRS08/09, Austroads, Sydney, NSW. Bali, S, Potts, R, Fee, JA, Taylor, JI & Glennon, J 1978, Cost effectiveness and safety of alternative roadway delineation treatments for rural two-lane highways: vol II: final report, FHWA/RD-78-51, Federal Highway Administration, Washington, DC, USA. Corben, B, Deery, H, Mullan, N & Dyte, D 1997, The general effectiveness of countermeasures for crashes into fixed roadside objects, report 111, Monash University Accident Research Centre, Clayton, Vic. Council, F & Stewart JR 1999, Safety effects of the conversion of rural two-lane to four-lane roadways based on cross-sectional models, Transportation Research Record, no. 1665, pp. 35-43. Council, F, McGee, HW, Prothe, L & Eccles, KA 2002, Run-off-road crash prevention in AASHTOs Strategic nd Highway Safety Plan, Institute of Transportation Engineers (ITE) Annual Meeting, 72 , 2002, Philadelphia, Pennsylvania, USA, Institute of Transportation Engineers (ITE), Washington, USA. Creasey, T & Agent, KR 1985, Development of accident reduction factors, research report UKTRP-85-6, Kentucky Transportation Research Program, University of Kentucky, Lexington, KY, USA. Elvik, R, Hye, A, Vaa, T & Srensen, M 2009, The handbook of road safety measures, 2 Publishing, Bingley, UK.
nd st

edn, Emerald

Ermer, DJ, Fricker, JD & Sinha,KC 1991, Accident reduction factors for Indiana, JHRP-91-11, Purdue University, School of Civil Engineering, Lafayette, IN, USA. Harwood, DW, Council, FM, Hauer, E, Hughes, WE & Vogt, A 2000, Prediction of the expected safety performance of rural two-lane highways, report FHWA-RD-99-207, Federal Highway Administration (FHWA), McLean, VA, USA. Hemion, RH 1969, A preliminary cost-benefit study of headlight glare reduction, report AR-683, Southwest Research Institute, San Antonio, TX, USA. Koorey, G, Farrelly, P, Mitchell, T & Nicholson, C 1999, Assessing passing opportunities: stage 2, research report 146, Transfund New Zealand, Wellington, NZ. Larsson, M, Candappa, N & Corben, B 2003, Flexible barrier systems along high speed roads: a lifesaving opportunity, report 210, Monash University Accident Research Centre, Clayton, Vic.

Austroads 2012 35

Effectiveness of Road Safety Engineering Treatments

McLean, J 1996, Review of accidents and rural cross section elements including roadsides, research report ARR 297, ARRB Transport Research, Vermont South, Vic. Miller, TR 1993, Benefit-cost analysis of lane marking, Public Roads, vol.56, no.4, pp.153-63. Moses, P 1990, Safety improved on rural road curves, Western Roads, vol.15, no.2, pp.2-3. Mutabazi, MI, Russell, ER & Stokes, RW 1999, Review of the effectiveness, location, design, and safety of passing lanes in Kansas, report K-TRAN:KSU-97-1, Kansas Department of Transportation, Topeka, KS, USA, viewed 25 October 2010, <http://ntl.bts.gov/lib/7000/7500/7576/784476.pdf>. Ogden, KW 1992, Benefit/cost analysis of road trauma countermeasures: rural road and traffic engineering programs, report 34, Monash University Accident Research Centre, Clayton, Vic. Parker, MR, Flak, MA, Tsuchiyama, KH, Wadenstorer, SC & Hutcherson, F 1983, Geometric treatments for reducing passing accidents at rural intersections on two lane highways: volumes 1 and 2, FHWA/RD83-074/75, Office of Safety and Traffic Operations, Federal Highway Administration, Washington, DC, USA. Road Safety Markings Association (RSMA) 2007, Road markings, road safety and efficient road utilisation in 21st century Britain, RSMA, Gainsborough, Lincolnshire, viewed 17 November 2010, <http://www.rsma.co.uk/files/whitelinessaveslives.pdf>. Slop, M & Catshoek, JWD 1995, Recommended safety measures for application on interurban roads in the short term, report R-95-18, Institute for Road Safety Research, Leidschendam, Netherlands. Thrush, M 1996, Assessing passing opportunities: literature review, research report 60, Transit New Zealand, Wellington. VicRoads 1990, Guidelines for the selection of projects under the road conditions sub-program (incorporating accident blackspot projects, mass action projects, railway level crossing projects), Road Safety Division, VicRoads, Kew, Vic.

Austroads 2012 36

Effectiveness of Road Safety Engineering Treatments

C.4
C.4.1

Profile Line Marking


Profile Edge Lining, Edge Line (Shoulder) Rumble Strip, or Shoulder Grooving
Study Year 1982 1985 Country United States United States Environment/Treatment Installation of profile edge lines on rural highway Installation of rumble strips on various State roads (edge line or centreline not specified) Motorway left-hand bends Installation of profile edge lines on rural highway Installation of profile edge lines on rural highway Installation of shoulder rumble strips motorway Shoulder rumble strips Rumble strips Installation of audio-tactile edge line Reduction 61% (run-off-road) 25% (total crashes)

Summary of research

FHWA in Ogden (1996) Creasy and Agent

UK Department of Highways and Planning in Cairney (1993) Anon in Ogden (1996) Harwood in Ogden (1996) County Surveyors Society in Ogden (1996) Wood Harwood RTA

1988

UK

76% (run-off-road)

1988 1993 1989 1994 1995 1995

United States United States UK United States United States Australia

49% (run-off-road),19% (total crashes) 20% or more (run-off-road) 37% (total crashes, not significant) 76% (run-off-road, statistically significant) 70% (run-off-road) 50% (run-off-road) Single vehicle crashes 20% (permanent obstruction in carriageway) 30% (off carriageway on straight) 30% (off carriageway on straight into object) 30% (off carriageway on curve) 30% (off carriageway on curve into object) 25%

Agent, Stamatiadis & Jones

1996

United States

Installation of rumble strips on various State roads (edge line or centreline not specified) Installation of shoulder grooving on various State roads Installation of audio-tactile edge line Installation of audible tactile edge lines Pennsylvania turnpike. Treatment includes RRPMs however authors did not expect large contribution from recessed reflective PMs

Agent, Stamatiadis & Jones Corben et al. Cairney Hickey

1996

United States

25%

1996 1996 1997

Australia Australia United States

9% (total crashes) 33% (total crashes) 64% (run-off-road)

Austroads 2012 37

Effectiveness of Road Safety Engineering Treatments

Study Perrillo

Year 1998

Country United States

Environment/Treatment Installation of shoulder rumble strips on New York thruway Installation of shoulder rumble strips The report identifies this treatment as rumble strips, however this could be transverse lines on approach to intersection considering crash mix Installation of shoulder grooving Unknown road environment (edge line or centreline not specified) Pavement shoulder grooving/strip Groove shoulder

Reduction 72% (run-off-road)

Baas, Charlton & de Jong Shen et al.

2001

New Zealand

27% (total crashes) 32% (run-off-road) 42% (fatal crashes) 40% for all crashes 76% for head-on crashes 50% for rear-end crashes 54% for right-angle crashes 100% for side-swipe crashes 33% for left-turn crashes 51% for run-off-road crashes 22% (total crashes) 27% (run-off-road) Average 32% (all) Average 80% (head-on) Average 51% (run-off-road) 22% (all) 27% (run-off-road) 25% (all) 18% (injury) 17% (PDO) 27% (run-off-road)

2004

United States

Shen et al. Gan, Shen & Rodriguez

2004 2005

United States United States

Gan, Shen & Rodriguez Gan, Shen & Rodriguez

2005 2005

United States United States

Griffith in Bahar et al. (2007)

1999

United States

Installing edge line rumble strips

Run-off-road type crashes 13% (injury, all freeways) std error: 12% 21% (all types, rural, freeway) std error: 10% 18% (all types, all freeways) std error: 7% Run-off-road type crashes 34% (all types, rural, all roads) 16% (all types, rural, arterial roads) 34% (all types, rural, between ramps) 38% (all types, rural, highways) 36% (all types, rural, 3-lane roads) 32% (all types, rural, 2-lane roads) 14% (all crashes) 47% (night/wet pavement crashes)

Smith et al. in Bahar et al. (2007)

2005

United States

Installing edge line rumble strips

NYSDPOT

2009

United States

Intra-state before and after evaluation of shoulder rumble strips installed on the shoulder of high volume roads Summary of various studies regarding the effect of installing shoulder rumble strips on roads

Elvik et al.

2009

Various

10% (total crashes) 25% (single vehicle crashes) 16% (run-off-road, all severities controlled for publication bias) 52% (run-off-road, injury crashes) 44% (run-off-road, all severities not controlled for publication bias)

Austroads 2012 38

Effectiveness of Road Safety Engineering Treatments

Study ADOT

Year 2009

Country United States

Environment/Treatment Before and after evaluation of roadside countermeasures including the installation of rumble strips in the State of Arizona (edge line or centreline not specified)

Reduction All crash types: 53% (all casualties), 83% (fatalities), 65% (injuries) Run-off-road crashes: 54% (all casualties), 75% (fatalities), 56% (injuries) Side-swipe crashes and head-on crashes: 80% (all casualties) All crash types: 18% (all casualties), 15% (fatalities), 18% (injuries) run-off-road crashes: 27% (all casualties), 12% (fatalities), 27% (injuries)

ADOT

2009

United States

Shoulder grooving

AASHTO

2010

United States

Unknown study on milledin shoulder rumble strips on rural multilane divided highways and installing continuous rumble strips on freeway shoulders

Rural multi-lane divided 16% (all types, all severities) std error: 10% 17% (all types, injury crashes) std error: 20% 10% (run-off-road, all severities) std error: 30% 12% (run-off-road, injury crashes) std error: 30% Freeways 79% (run-off-road, urban/rural freeway, all crash severities) std error: 7% specific subset of run-off-road crashes) 18% (run-off-road, urban/rural freeway, all crash severities) std error: 10% 13% (run-off-road, urban/rural freeway, all injury crashes) std error: 20% 21% (run-off-road, rural freeway, all severities) std error: 20% 7% (run-off-road, rural freeway, injury crashes) std error: 30% 23% (all crash types)

Austroads

2010

Australia

Profile edge lining, or shoulder grooving

Thirteen studies provided crash reduction factors associated with profile edge lines for all crashes. Injury crashes were used where available. The average reduction in casualty crashes identified as a result of the installation of profile edge lines (or shoulder grooving/rumble strips) was 21% (rounded to 20%). There is a medium level of confidence, as while there was a range in the reductions, many of the studies identified reductions that were similar to the estimated value. Nineteen studies provided crash reductions for run-off-road type crashes; the average of these was 39% (rounded to 40%). This figure was assessed as of a low level of confidence due to the wide spread of crash reduction factors. It was also noted that a number of studies provided crash reductions for a number of road environments (Bahar et al. 2007 & AASHTO 2010).

Austroads 2012 39

Effectiveness of Road Safety Engineering Treatments

C.4.2

Profile Centre Lining, Centreline Rumble Strip


Study Year 1995 Country United States Environment/Treatment Installation of centreline rumble strips on two-lane rural highways Installation of audio-tactile centreline Installation of centreline rumble strips Installation of centre rumble strips on two-lane mountain highways Installation of centreline rumble strips Installation of centreline rumble strips on undivided two-lane roads Installing centreline rumble strips Reduction 39% (head-on crashes) 100% (fatal head-on crashes) 14% (injury head-on crashes) 68% (PDO crashes) 10% (off carriageway on curve) 10% (off carriageway on curve into object) 58% (head-on crashes) 34% (head-on crashes) 36% (side-swipe crashes) 2137% (head-on and side-swipe crashes) 14% (total crashes) 25% (head-on and side-swipe crashes) 22% (injury, rural, multi-lane divided) before/After study) 10% (all types, rural, multi-lane divided) before/After study) 18% (fatal/injury, rural, 2-lane roads) standard error: 12% 13% (all types, rural, 2-lane roads) standard error: 8% 4% (total crashes) (CI:19%, 13%) 24% (head-on, injury crashes) (33%, 13%)

Delaware State (in AASHTO n.d.)

RTA Fitzpatrick et al. Outcalt

1995 2000 2001

Australia United States United States

Baas, Charlton & de Jong Persaud, Retting, & Lyon

2001 2003

New Zealand Canada

Carrasco et al. in Bahar et al. (2007)

2004

United States

Patel et al. in Bahar et al. (2007)

2007

United States

Installing centreline rumble strips

Elvik et al.

2009

Various

Summary of various studies regarding the effect of installing centreline rumble strips on roads Centreline rumble strips for rural two-lane roads

AASHTO

2010

United States

14% (all types, all severities) std error: 5% 15% (all types, injury crashes) std error: 8% 21% (head-on & opposing direction side-swipe, all severities) std error: 10% 25% (head-on & opposing direction side-swipe, injury crashes) std error: 20%

Five studies provided general crash reductions for the provision of profile centrelines. Injury crashes were used where available. The effectiveness of profile centrelines in reducing crashes of all types was found to be 15% (medium level of confidence).

Austroads 2012 40

Effectiveness of Road Safety Engineering Treatments

Six studies provided crash reductions for head-on type crashes. The average reduction was 30%, which is within the effectiveness range identified by Baas, Charlton and de Jong (2001). The figure was assessed as having a low level of confidence due to the wide spread of crash reduction factors in the literature reviewed. Note that the Persaud, Retting and Lyon (2003) crash reduction factor for head-on and side-swipe crashes was considered a head-on crash type for this study because other studies from North America discuss head-on and opposing direction side-swipe which is head-on in the Australian context. Further, inclusion of this figure did not change the identified crash reduction figure. References assessed
AASHTO n.d., Head on collisions: description of strategies, American Association of State Highway and Transportation Officials, Washington, DC, USA, viewed 7 February 2006, <http://safety.transportation.org/htmlguides/HOcrashes/description_of_strat.htm> AASHTO 2010, Highway safety manual, 1 edn, American Association of State Highway and Transportation Officials, Washington, DC, USA. ADOT 2009, Benefit/cost ratio economic analysis, section 231, in Traffic engineering policies, guides and procedures (PGP), Arizona Department of Transportation, Phoenix, AZ. viewed 25 July 2011, < http://www.azdot.gov/highways/Traffic/standards/PGP/TM231.pdf>. Agent, KR, Stamatiadis, N & Jones, S 1996, Development of accident reduction factors, research report KTC-96-13, Kentucky Transportation Centre, University of Kentucky, Lexington, KY, USA. Austroads 2010, Road safety engineering risk assessment: part 6: crash reduction rates, AP-T151-10, Austroads, Sydney, NSW. Baas, P, Charlton, S & de Jong D 2001, Review of lane delineation, Transport Engineering Research NZ, Maukau City, New Zealand. Bahar, G, Masliah, M, Wolff, R & Park, P 2007, Desktop reference for crash reduction factors, report FHWASA-07-015, Federal Highway Administration, Washington DC, USA. Cairney, P 1993, Current issues in delineation, Road and Transport Research, vol.2, no.2, pp.28-39. Cairney, P 1996, Evaluation of audible edge lines , contract report CR C5567, ARRB Transport Research Ltd, Vermont South, Victoria, Australia. Corben, B Newstead, S Diamantopoulou, K & Cameron, M 1996, Results of an evaluation of TAC funded accident black spot treatments. Combined 18th ARRB Transport Research Conference and Transit New Zealand Land Transport Symposium, 1996, Christchurch, New Zealand. ARRB Transport Research Ltd, Vermont South, Victoria, Australia. Creasey, T & Agent, KR 1985, Development of accident reduction factors, research report UKTRP-85-6, Kentucky Transportation Research Program, University of Kentucky, Lexington, KY, USA. Elvik, R, Hye, A, Vaa, T & Srensen, M 2009, The handbook of road safety measures, 2 Publishing, Bingley, UK.
nd st

edn, Emerald

Fitzpatrick, K, Balke, K Harwood DW & Anderson, IB 2000, Accident mitigation guide for congested rural two-lane highways. NCHRP report 440, Transportation Research Board of the National Academies, Washington, D.C.

Austroads 2012 41

Effectiveness of Road Safety Engineering Treatments

Gan, A, Shen, J & Rodriguez, A 2005, Update of Florida crash reduction factors and countermeasures to improve the development of district safety improvement projects, Lehman Centre for Transportation Research, Miami, Florida, USA. Harwood, D W 1995, Enhancing highway safety with rumble strips, TR News, 178, pp. 12-16. Hickey, JJ 1997, Shoulder rumble strip effectiveness: drift-off-road accident reductions on the Pennsylvania Turnpike, Transportation Research Record, no. 1573, pp.105-9. New York State Department of Transport 2009, PIES: reduction factor report, NYSDOT, Albany, New York. Outcalt, W 2001, Centerline Rumble Strips, report CDOT-DTD-R-2001-8, Colorado Department of Transportation, Denver, Colorado. Ogden, KW 1996, Safer roads: a guide to road safety engineering, Avebury Technical, Aldershot, UK. Perrillo, K 1998, The effectiveness and use of continuous shoulder rumble strips, Federal Highway Administration, Albany, New York. Persaud, BN, Retting, RA & Lyon, C 2003, Crash reduction following installation of centerline rumble strips on rural two-lane roads, Ryerson University, Toronto, Canada. RTA NSW 1995, Accident investigation and prevention: policy and guidelines; version 1.0, no: 95.090, Roads and Traffic Authority New South Wales, Sydney, NSW. Shen, J Rodriguez, A Gan, A & Brady, P 2004, Development and application of crash reduction factors: A state-of-the-practice survey of State Departments of Transportation, Transportation Research Board annual meeting, 83rd, Washington DC, TRB, Washington, DC, USA. Wood, NE 1994 Shoulder rumble strips: a method to alert drifting drivers, Pennsylvania Turnpike Commission, Harrisburg, Pennsylvania.

Austroads 2012 42

Effectiveness of Road Safety Engineering Treatments

C.5

Signs Regulatory
Study Year 1985 1982 1995 1991 Country US Australia UK Indiana, US Environment/Treatment All regulatory signs at intersections Regulatory signs at intersections Install mandatory signs Sign installation (regulatory signs not specified) 50% 38% 55% CRF 0.15 Reduction

Summary of research
General Creasey and Agent Moses Mountain, Jarret & Fawaz Ermer, Fricker & Sinha

Install stop sign Andreassen Ewing Shen et al. Elvik et al. 1989, but based on 1962 study 1999 2004 2009 Australia US US Various Stop sign Stop Stop sign Stop at three-leg intersection Stop two-way at fourleg intersection Kulmala, R 1994 Finland Stop sign: Three-leg intersection Four-leg intersection Agent et al. Gan, Shen & Rodriguez El-Basyouny & Sayed Guyano-Cardona, Sylvester & Jenkins Agent, Stamatiadis & Jones Bahar et al. Elvik et al. Bahar et al. 1996 2005 2010 2002 1996 2007 2009 2007 US US Canada Australia US Various (mainly US) Various Various (mainly US) Two-way stop Two-way stop Stop two-way at fourleg intersection Four-way stop in Australia Four -way stop Double stop sign (presume two-way stop) Four -way stop Two-way to four-way stop 46% all crashes, 9% injury crashes 2% all crashes, 12% injury crashes 35% all crashes 35% 51% 83% 55% all crashes All 11% Right angle 3655% 45% All: 4764% Urban injury 71% Urban right-turn 20% Pedestrians 39% Rear-end 13% Right angle 1547% 60% 70% 35% 19% injury crashes 35% injury crashes

Austroads 2012 43

Effectiveness of Road Safety Engineering Treatments

Study Gan, Shen & Rodriguez

Year 2005

Country US

Environment/Treatment Two-way to four-way stop All: 53% Rear-end: 13% Right angle 72% Right-turn 20% Pedestrians 39%

Reduction

Harkey et al.

2008

US

Combination to four-way stop

All crashes: 0.53 CMF Angle: 0.28 CMF Rear-end: 0.87 Right-turn: 0.8 Pedestrian: 0.61 Injury: 0.29 CMF CMF 0.3 (urban) CMF 0.52 (rural) but all severities 45% all crashes 3% injury crashes 3% property damage only crashes 29% 31% Right angle 43% 11% increase All: 137% increase Urban: 127% increase All: 2.37 CMF 39% increase

AASHTO Install Give-way sign Agent, Stamatiadis & Jones Elvik et al. Shen et al. Gan, Shen & Rodriguez Stop sign to Give way sign Pegrum, Lloyd & Willett Bahar et al. Harkey et al. Elvik et al. Give way sign to Stop sign Bahar et al. Right-turn and U-turn ban Brich & Cottrell Austroads (based on the revised tables for FORS, 1996 prepared by Dr David Andreassen for the Australian Transport Safety Bureau)

2010

US

Two-way to four-way stop

1996 2009 2004 2005

US Various US US

Give way sign Give Way sign Give way sign Give way sign

1972 2007 2008 2009

Australia Various (mainly US) US Various

Stop sign to give way undivided Stop sign to give way sign Stop sign to give way sign Replace stop with give way sign Replace give way with stop sign No U-turn, ban right-turn at signalised intersections Ban right-turn

2007

Various (mainly US) US Australia

All 29% Right angle 9% (left angle Australia) 66% 50% reduction in opposing turns, parallel lanes turning and loss of control crashes 50% reduction in rear-end, parallel lanes turning and loss of control crashes 50% reduction in U-turn, rear-end, parallel lanes turning and loss of control crashes

1995 2009

Ban left-turn

Ban U-turn

Austroads 2012 44

Effectiveness of Road Safety Engineering Treatments

Study Bahar et al.

Year 2007

Country Various (mainly US)

Environment/Treatment No U-turn & ban right-turn

Reduction All: 62% Right or U-turn: 59% All: 45% Right-turn: 90% Pedestrian: 10% Rear-end: 30% CMF 0.32 (all intersection crashes, all severities), right-turn CMF 0.36

Right-turn ban

AASHTO

2010

US

Right-turn ban at stop & signalised intersections, and median crossovers No U-turn & ban right-turn

CMF 0.28 (all intersection crashes, all severities), right-turn CMF 0.23 Dependent on effect diverted traffic will have on other routes

Neuman et al.

2003

US

Turn restriction or prohibition

C.5.1

General: Regulatory Signs at Intersections

Three studies provided a general crash reduction for the installation of regulatory signs at intersections:

Creasy and Agent: 50% Moses: 38% Mountain, Jarret and Fawaz: 55%.

It is recommended that a general reduction not be used, as there is generally sufficient information on stop and give way signs to provide separate figures for each of these, which would be more useful for practitioners. These figures also seem high compared to reductions associated with stop and give way signs individually. Note, the Andreassan (1989) figure for installation of a stop sign has not been included due to the age of the study (based on 1962 data). Ewing (1999) and Shen et al. (2004) provide general reductions for installation of a stop sign. These figures have not been included as there is a large variation in results for these studies, and there is generally sufficient information on stop signs at three and four-leg intersections to provide separate figures for each of these. Haleem, Abdel-Aty and Mackie (2010) note that stop signs (rather than give way signs) tend to be installed at locations where crashes are more likely to happen. It should be noted that selection of stop or give way signs for a given location is generally dictated by relevant warrants, which relate to sight distance, rather than selecting them based on potential crash reductions. It is important that regulatory signs are installed in accordance with the relevant warrants, in order to provide a consistent approach to intersection management for drivers. C.5.2 Stop Signs at Three-leg Intersections

Only two studies were found that provided figures for provision of stop signs at three-leg intersections (Elvik et al. 2009 19% injury; Kulmala 1994 9% injury). Based on these, a reduction of 14% is proposed (rounded to 15%). Low confidence is placed in this estimate as it uses only two studies and there is quite a large variation in results from those studies.

Austroads 2012 45

Effectiveness of Road Safety Engineering Treatments

C.5.3

Stop Signs at Four-leg Intersections Kulmala: 12% injury (2% all) Elvik et al.: 35% Agent, Stamatiadis and Jones: 35% Gan, Shen and Rodriguez: 35% El-Basyouny: 51.5% Bahar et al.: 11% (angle crashes 3655%).

Six studies provided crash reductions for installation of stop signs at four-leg intersections:

Based on these, an average crash reduction of 30% is proposed. There is medium confidence in this figure, as it is based a number of studies, although there is quite a large variation in some results. C.5.4 Four-way Stop Signs

Three studies provided crash reductions for installation of four-way stop signs (Guyano-Cardona 83%, Agent, Stamatiadis & Jones 55%, Elvik et al. 45%) which provided an average crash reduction of 61% (rounded to 60%, low confidence). Similar results were found for conversion of two-way to four-way stops. Interestingly, AASHTO reported larger crash reductions for urban environments. C.5.5

Give-way Signs Agent, Stamatiadis and Jones: 45% Elvik et al.: 3% Shen et al.: 29% Gan, Shen and Rodriguez: 31%.

Four studies provided crash reductions for installation of give-way signs:

The reductions from these studies vary considerably. The average of these is 27% (rounded to 25%). However, it is unclear whether this figure applies to three-leg or four-leg intersections or both. It is noted that for the installation of stop signs, the crash reduction is considerably lower at three-leg compared to four-leg intersections. There is low confidence in this figure. C.5.6

Stop Sign to Give-way Sign Pegrum, Lloyd and Willett: 11% increase Bahar et al.: 137% increase Harkey et al.: same as Bahar et al. Elvik et al.: 39% increase.

Four studies provided figures for conversion of stop to give-way control:

The figures in each of the studies were very different, and therefore the information is not considered reliable enough to provide a recommended crash reduction.

Austroads 2012 46

Effectiveness of Road Safety Engineering Treatments

C.5.7

Give-way Sign to Stop Sign

One study provided a crash reduction for the situation of replacing a give-way sign with a stop sign. Due to lack of data, an associated crash reduction is not recommended. C.5.8 U-turn and Right-turn Ban

A number of studies provide crash reductions for right-turn bans, or U-turn and right-turn bans, with crash reductions ranging from 45% to 72%. Given the limited studies available, and uncertainty in crash reductions, an overall figure of 61% is recommended (rounded to 60%, medium confidence). As observed by Neuman et al. (2003), the actual crash reduction experienced will be dependent on the effect that diverted traffic has on nearby routes and intersections, and therefore consideration needs to be given to the potential for crash migration. References assessed
AASHTO 2010, Highway safety manual, 1 edn, American Association of State Highway and Transportation Officials, Washington, DC, USA. Agent, KR, Stamatiadis, N & Jones, S 1996, Development of accident reduction factors, research report KTC-96-13, Kentucky Transportation Centre, University of Kentucky, Lexington, KY, USA. Andreassen, DC 1989, Strategies for safety problems, research report ARR 163, Australian Road Research Board, Vermont South, Vic. Andrew OBrien and Associates 2000, Managing traffic flow on urban freeways: appendix A: literature review, Andrew OBrien and Associates, Melbourne, Vic. Austroads 2009, Guide to road safety: part 8: treatment of crash locations, AGRS08/09, Austroads, Sydney, NSW. Bahar, G, Masliah, M, Wolff, R & Park, P 2007, Desktop reference for crash reduction factors, report FHWASA-07-015, Federal Highway Administration, Washington DC, USA. Brich, SC & Cottrell, BH Jnr 1995, Guidelines for the use of no u-turn and no left-turn signs, report VTRC 95-R5, Virginia Transportation Research Council, Charlottesville, Virginia. Bureau of Transport Economics (BTE) 2001, The black spot program 1996-2002: An evaluation of the first three years, report no. 104, Bureau of Transport Economics (BTE), Canberra, Australia. Cairney, P 1984, Casualty accidents in the vicinity of traffic control devices, Ergonomics society conference, 1984, Exeter, UK, Taylor and Francis, pp. 210-215. Creasey, T & Agent, KR 1985, Development of accident reduction factors, research report UKTRP-85-6, Kentucky Transportation Research Program, University of Kentucky, Lexington, KY, USA. El-Basyouny, K & Sayed, T 2010, A full bayes approach to before-after safety evaluation with matched th comparison, 89 annual meeting of the Transportation Research Board, TRB, Washington, DC, USA. Elvik, R, Hye, A, Vaa, T & Srensen, M 2009, The handbook of road safety measures, 2 edn, Emerald Publishing, Bingley, UK. Ermer, DJ, Fricker, JD & Sinha, KC 1991, Accident reduction factors for Indiana, JHRP-91-11, Purdue University, School of Civil Engineering, Lafayette, IN, USA.
nd st

Austroads 2012 47

Effectiveness of Road Safety Engineering Treatments

Ewing, R 1999, Traffic calming: state of the practice, report FHWA-RD-99-135, Federal Highway Administration, Washington, DC, USA. Gan, A, Shen, J & Rodriguez, A 2005, Update of Florida crash reduction factors and countermeasures to improve the development of district safety improvement projects, Lehman Centre for Transportation Research, Miami, Florida, USA. Guyano-Cardona, J, Sylvester, P & Jenkins, I 2002, Four-way stop signs: the Newcastle experience, Travelator, Proceedings of the 2002 AITPM National Conference, AITPM, Thornleigh, NSW. Haleem, K, Abdel-Aty, M & Mackie, K 2010, Using a reliability process to reduce uncertainty in predicting crashes at unsignalized intersections, Accident Analysis and Prevention, vol. 42, no.2, pp. 654-66. Harkey, DL, Srinivasan, R, Baek, J, Council, FM, Eccles, K, Lefler, N, Gross, F, Persaud, B, Lyon, C, Hauer, E & Bonneson, JA 2008, Accident Modification Factors for Traffic Engineering and ITS Improvements, NCHRP report 617, Transportation Research Board, Washington, DC, USA. Kulmala, R 1994, Measuring the safety effect of road measures at junctions, Accident Analysis and Prevention, vol. 26, no. 6, pp. 781-94. Moses, P 1982, Traffic signs evaluated, Western Roads, Main Roads Western Australia, vol.7, no.3, 7pp. Mountain, L, Jarret, D & Fawaz, B 1995, The safety effects of highway engineering schemes, Proceedings of the Institution of Civil Engineers: Transport, vol.111, no.4, pp.298-309. Neuman TR, Pfefer R, Slack KL, Kennedy Hardy K, Harwood DW, Potts IB, Torbic DJ, Kohlman Rabbani ER 2003, Guidance for implementation of the AASHTO strategic highway safety plan: volume 5: A guide for addressing unsignalized intersection collisions, NCHRP report 500, Transportation Research Board, Washington, DC, USA. Ogden, KW 1996, Safer roads: a guide to road safety engineering, Avebury Technical, Aldershot, UK. Pegrum, BV, Lloyd, ER & Willett, P 1972, Experience with priority roads in the Perth metropolitan area, th Australian Road Research Board conference, 6 , Canberra, Australian Road Research Board, Vermont South, Vic, vol.6, no.3, pp.363-383. Shen, J Rodriguez, A Gan, A & Brady, P 2004, Development and application of crash reduction factors: A state-of-the-practice survey of State Departments of Transportation, Transportation Research Board annual meeting, 83rd, Washington DC, TRB, Washington, DC, USA.

C.6

Traffic Signals

Summary of research Due to the volume of information on this topic, the summary tables have been split into provision of new signals, and the effect of turn phases. C.6.1 New Signals
Study Corben, Ambrose & Chee Wai Year 1990 Country Australia Environment/Treatment Blackspots 53% Reduction

Austroads 2012 48

Effectiveness of Road Safety Engineering Treatments

Study Ogden (referenced 5 studies including Corben 1990)

Year 1996

Country Australia, UK, US

Environment/Treatment New signals

Reduction 15 53% (overall) (1 instance with site with low crash numbers resulted in 5.3% increase) Angle crashes: 52 84% Rear-end: 31% reduction to 65% Turn opposing direction: 52% 75% General: 25% Angle crashes: 65% 32% 47.1% 42.9% 76.0% 37.3% 27% 20% All crashes: 14% Right-angle crashes: 34% Rear-end crashes: 50% increase All crashes: 27% Right-angle crashes: 67% Rear-end crashes: 38% increase All crashes: 0.58% increase in crashes Rear-end: 81.81% increase Angle: 29.02% reduction Left-turn: 28.24% reduction Right-turn: 50.2% reduction Side-swipe: 28.57% increase Pedestrian: 17.4% reduction All other: 131.7% increase Fatal: 38% reduction Injury: 5% reduction No injury: 14.8% increase Rural: 35.03% reduction Urban: 16.6% increase

Agent, Stamatiadis & Jones Newstead & Corben BTE

1996 2001 2001

US Australia Australia

Safety improvement programs Blackspots Blackspots New signals, no arrows (urban) New signals, with arrows (urban) New signals, no arrows (rural) New signals, with arrows (rural) New traffic signals New signals + turn lane Urban three-leg intersections

Thomas & Smith Persaud et al.

2001 2003

US US

Urban four-leg intersections Pernia et al. 2004 US Rural and urban

Gan, Shen & Rodriguez

2005

US

New signals

All: 22% Rear-end: 21% Left angle: 58% Overturn: 21% All:36% Injury: 53% Rear-end: 8% Left angle: 74% 35.0% 49.9% 26.6% 33.1%

New signals from two-way stop

Scully et al.

2006

Australia

All signal treatments New signals Fully controlled right turn Mast arm

Austroads 2012 49

Effectiveness of Road Safety Engineering Treatments

Study DIER

Year 2007

Country Australia

Environment/Treatment Blackspots (only 2 sites)

Reduction 71% reduction in casualty crashes Adjacent approaches: 100% Rear-end: 22% Pedestrian: 22% <5000 veh/lane/day: 38% >5000 veh/lane/day: 20 Rural: 15% Urban: 17% All severities: 2038% Fatal crashes: 38% Urban casualty three-leg: 14% Urban casualty four-leg: 23% Rural: 15% Urban three-leg: all crashes: 0.86 CMF, angle 0.66 CMF, rear-end 1.5 CMF Urban four-leg: all crashes: 0.77 CMF, angle 0.33 CMF, rear-end 1.38 CMF Rural: all crashes: 0.56 CMF, angle 0.23 CMF, rear-end 1.58 CMF, right-turn 0.4 CMF

Bahar et al.

2007

US

Not given

FHWA (includes references to a number of others already referenced) Harkey et al.

2009

US

Install signals

2008

US

Install signals

Meuleners et al. Elvik et al.

2008 2009

Australia International

Install signals Not given

21.2% signalising a three-leg intersection: 15% (95% Cl: 25; 5) signalising a four-leg intersection: 30% (95% Cl: 35; 25) Install signals All crash types, all severities CMF 0.95 (standard error 0.09) Right-angle crashes, all severities CMF 0.33 (standard error 0.06) Rear-end crashes, all severities 2.43 (standard error 0.4)

AASHTO

2010

US

Urban, speed limit at least 40 mph, four-leg intersections

AASHTO

2010

US

Rural, three and four-leg intersections

Install signal All crash types, all severities CMF 0.56 (standard error 0.03) Right angle crashes, all severities CMF 0.23 (standard error 0.02) Left-turn crashes, all severities CMF 0.40 (standard error 0.06) Rear-end crashes, all severities 1.58 (standard error 0.2)

In relation to all crashes, seven studies provided general crash reductions for installation of traffic signals. However, the figure from Tasmania was based on only two sites, and appeared to be an outlier. Therefore it was omitted. Based on the remaining six studies, a reduction of 30% is recommended. Due to the variability in results, there is a low level of confidence in this figure.

Austroads 2012 50

Effectiveness of Road Safety Engineering Treatments

In relation to crash types, three studies provided figures for reductions in angle crashes and two studies provided results for rear-end crashes. Studies showed mixed results for rear-end crashes, with many showing increases in this crash type. Based on these results a general reduction of 50% for angle crashes, and a 30% increase in rear-end crashes are recommended. Installation of signals is likely to result in an increase in rear-end crashes, but reduce angle and pedestrian-type crashes, which are often higher severity. A number of studies provided figures for installing signals at three and four-leg intersections. There appears to be greater benefit for signalisation of four-leg intersections, which is not surprising as movements at four-leg intersections are generally more complicated than at three-leg intersections. A number of studies also provided figures for the provision of signals at urban and rural intersections. Based on these, there are some mixed results, however there appears to be great improvement for rural installations. C.6.2 Effect of Turn Phases
Study Corben, Ambrose & Chee Wai Hall in Wilke & Appleton (2005) Year 1990 1993 Country Australia New Zealand Environment/Treatment Blackspots Lag right-turn (filter rightturns, followed by rightturn arrow) Lead right-turn (right-turn arrow, followed by filter right-turns) Fully controlled right-turn (no filtering) None to partially controlled right-turn Reduction Right-turn phase installation 44% 30% 68% 90% All: no significant effect (casualty) Right through: 18% increase Cross traffic, pedestrian, rear-end & left rear: no significant effect Right rear: 85% reduction Other: 32% reduction All: 45% reduction (casualty) Right through: 82% reduction Cross traffic: 48% reduction Pedestrian: 35% reduction Right rear: no significant effect Rear-end & left rear: 72% increase Other: 33% reduction All: 65% reduction (casualty) Right through: 93% reduction Cross traffic: 51% reduction Pedestrian, right rear, rear-end & left rear, other: no significant effect 3373% (general) 82% turn opposite direction 48% right angle 35% pedestrian 72% increase rear-end

Bui, Cameron & Foong

1991

Australia

Bui, Cameron & Foong

1991

Australia

None to fully controlled right-turn

Bui, Cameron & Foong

1991

Australia

Partially controlled to fully controlled right-turn

Ogden (referenced 3 studies)

1996

Australia, UK

Fully controlled right-turn

Austroads 2012 51

Effectiveness of Road Safety Engineering Treatments

Study Agent, Stamatiadis & Jones

Year 1996

Country US

Environment/Treatment Safety improvement programs

Reduction General: 25% Exclusive right-turn phase; 25% (70% of right-turn crashes) Protected/permissive right-turn phase: 10% (40% of right-turn crashes) 32% 36% 58% All crashes: 27%, rear-end 31%, left angle 68%, right-turn 48%, overturn 31% All crashes: 10%, right-turn 40% Protected right-turn phase: 16% right-turn 19% angle Protected/permissive right-turn phase: 16% of right-turn crashes, 12% left-angle crashes (casualty crashes) Leading green arrow: 17% of right-turn crashes, 25% left-angle crashes (casualty crashes) Urban casualties: 16% right-turn, 19% left-angle All severities: 2736% Right-turn: 4148% Left angle: 5463% Rear-end: 2735%

Newstead & Corben Thomas & Smith

2001 2001

Australia US

Blackspots Fully controlled right-turn Add phasing to existing signals Add phasing + turn lane(s) Exclusive right-turn Protected/permissive right-turn

Gan, Shen & Rodriguez

2005

US

Lyon et al. (in Bahar et al. 2007) Bahar et al.

2005

US

Urban

2007

US

Urban

FHWA (includes references to a number of others already referenced)

2009

US

Protected right-turn

Elvik et al.

2009

International

Right-turn crashes

Establishing right-turn phase No severity info: CRF 10% (95% CI -15;-5) Elvik text = many of the results given are very uncertain and should be treated with scepticism Separate right-turn phase (Fully controlled right turn) No severity info: CRF 58% (95% CI -64;-50) Elvik text = many of the results given are very uncertain and should be treated with scepticism Conflict-free phase changes Injuries CRF 75% (95% CI -90;-35) Elvik text = many of the results given are very uncertain and should be treated with scepticism

Elvik et al.

2009

International

Right-turn crashes

Elvik et al.

2009

International

AASHTO

2010

US

Urban three and four-leg signalised intersections Change to protected (controlled) phasing

Left-turn crashes all severities: CMF 0.01 (standard error 0.01) All types of crashes all severities: CMF 0.94 (standard error 0.1)

Austroads 2012 52

Effectiveness of Road Safety Engineering Treatments

Study AASHTO

Year 2010

Country US

Environment/Treatment Urban four-leg signalised (major road 3000 to 77 000 v/p/d and minor road 1 to 45 500) Urban four-leg signalised unspecified AADT

Reduction Change from filtered to semi-controlled For left-turn injuries CMF 0.84 (standard error 0.02) Change from filtered to semi-controlled For all crash types and all severities CMF 0.99 (standard error unknown)

AASHTO

2010

US

One study considered the provision of fully controlled right-turns, provision of partial control and changing partial to full control (Bui et al. 1991). The study (p.25) observed that full control of right-turns yields poorer intersection performance than partial control under virtually all conditions. The differences in performance, particularly between partial and full control, were slight and are unlikely to negate the safety advantages. Ten studies provided figures for the provision of fully controlled right-turns. Five studies provided general reductions for the provision of fully controlled right-turns (Corben et al. 1990, Bui, Cameron & Foong 1991, Agent, Stamatiadis & Jones 1996, Newstead & Corben 2001 and Gan, Shen J & Rodriguez 2005). Based on these, a reduction of 35% is recommended (with a medium level of confidence). A number of studies also provided guidance for different crash types. Based on these, a reduction of 60% for right through crashes (low confidence) and 45% for adjacent direction crashes (low confidence) is recommended. There is mixed information regarding the effect on rear-end crashes, and so a recommended crash reduction has not been provided. Five studies provided figures for the provision of partially controlled right-turns. However, only one study provided a general figure of 10% (Agent, Stamatiadis & Jones 1996). There is low confidence in this figure as it is only based on one study. Two studies provided figures for changing partial to full control (Bui, Cameron & Foong 1991 and Elvik et al. 2009). Based on these, a general crash reduction of 70% is suggested. References assessed
AASHTO 2010, Highway safety manual, 1 edn, American Association of State Highway and Transportation Officials, Washington, DC, USA. Agent, KR, Stamatiadis, N & Jones, S 1996, Development of accident reduction factors, research report KTC-96-13, Kentucky Transportation Centre, University of Kentucky, Lexington, KY, USA. Bahar, G, Masliah, M, Wolff, R & Park, P 2007, Desktop reference for crash reduction factors, report FHWASA-07-015, Federal Highway Administration, Washington DC, USA. Bhesania, RP 1991, Impact of mast-mounted signal heads on accident reduction, ITE Journal, vol.61, no.10, pp.25-9. Bureau of Transport Economics (BTE) 2001, The black spot program 1996-2002: An evaluation of the first three years, report no. 104, Bureau of Transport Economics (BTE), Canberra, Australia. Bui, B, Cameron M & Foong, C 1991, Effect of right turn phases at signalised intersections: part 1: safety performance, report 20, Monash University Accident Research Centre, Clayton, Vic.
st

Austroads 2012 53

Effectiveness of Road Safety Engineering Treatments

Corben, BF, Ambrose, C & Chee Wai, F 1990, Evaluation of accident black spot treatments, report 11, Monash University Accident Research Centre, Clayton, Vic. DIER 2007, An evaluation of the national blackspot programme in Tasmania, Traffic Standards Branch, Department of Infrastructure, Energy and Resources, Hobart, Tasmania. Elvik, R, Hye, A, Vaa, T & Srensen, M 2009, The handbook of road safety measures, 2 edn, Emerald Publishing, Bingley, UK. Felipe, E, Mitic, D & Zein, SR 1998, Safety benefits of additional primary signal heads, Insurance Corporation Of British Columbia, Canada. FHWA & Institute of Transportation Engineers 2003, Making intersections safer : a toolbox of engineering countermeasures to reduce red-light running : an informational report, IR-115, Institute of Transportation Engineers, Washington, DC, USA. FHWA Office of Safety 2009a, Traffic signals, Intersection safety issue briefs, issue brief 5, FHWA-SA-10005 Federal Highway Administration Office of Safety, Washington, DC. FHWA Office of Safety 2009b, Retroreflective borders on traffic signal backplate: a South Carolina success story, FHWA-SA-09-011, Federal Highway Administration Office of Safety, Washington, DC. Gan, A, Shen, J & Rodriguez, A 2005, Update of Florida crash reduction factors and countermeasures to improve the development of district safety improvement projects, Lehman Centre for Transportation Research, Miami, Florida, USA. Harkey, DL, Srinivasan, R, Baek, J, Council, FM, Eccles, K, Lefler, N, Gross, F, Persaud, B, Lyon, C, Hauer, E & Bonneson, JA 2008, Accident Modification Factors for Traffic Engineering and ITS Improvements, NCHRP report 617, Transportation Research Board, Washington, DC, USA. Lyon, C, Haq, A, Persaud, BN & Kodama, ST 2005, Development of safety performance functions for signalized intersections in a large urban area and application to evaluation of left-turn priority treatment, Transportation Research Record, no. 1908, pp. 65-71. Meuleners, L, Hendrie, D, Legge, M & Cercarelli, LR 2005, An evaluation of the effectiveness of the black spot programs in Western Australia, 2000-2002, The University of Western Australia, Crawley, WA. Meuleners, L, Hendrie, D, Lee, AH & Legge, M 2008, Effectiveness of the black spot programs in Western Australia, Accident Analysis and Prevention, vol 40, no. 3, pp.1211-6. Newstead, S & Corben, B 2001, Evaluation of the 1992-1996 Transport Accident Commission funded accident black spot treatment program in Victoria, report 182, Monash University Accident Research Centre, Clayton, Vic. Ogden, KW 1996, Safer roads: a guide to road safety engineering, Avebury Technical, Aldershot, UK. Pernia, J, Lu, JJ, Zhuo, Y & Snyder, D 2004, Effects of traffic signal installation on intersection crashes, Advances in Transportation Studies, vol. 2, pp.83-96. Persaud, B, McGee, H, Lyon, C & Lord, D 2003, Development of a procedure for estimating expected safety effects of a contemplated traffic signal installation, Transportation Research Record, no. 1840, pp.96103.
nd

Austroads 2012 54

Effectiveness of Road Safety Engineering Treatments

Rodegerdts, LA, Nevers, B & Robinson, B 2004, Signalized intersections: informational guide, FHWA-HRT04-091, Federal Highway Administration (FHWA), Washington, DC, USA. Sayed, T, Leur, P & Pump, J 2005, Safety impact of increased traffic signal backboards conspicuity Transportation Research Board annual meeting, 84th, 2005, Washington, DC, USA, TRB, Washington, DC, USA. Scully J, Newstead S, Corben B, & Candappa N, 2006, Evaluation of the effectiveness of the $240M statewide blackspot program Accident blackspot component, Monash University Accident Research Centre, Clayton, Victoria, Australia Srinivasan, R, Council, FM, Lyon, C, Gross, F, Lefler, NX & Persaud, BN 2008, Safety effectiveness of selected treatments at urban signalized intersections, Transportation Research Record, no. 2056, pp.70-6. Thomas, GB & Smith, DJ 2001, Effectiveness of roadway safety improvements, final report, Center for Transportation Research and Education, Iowa State University, USA. Turner, S, Turner, B & Wood, G 2008, Accident prediction models for traffic signals, ARRB conference, 23rd, 2008, Adelaide, South Australia, Australia, ARRB Group Ltd, Vermont South, Vic, 20pp. Wilke, A & Appleton, I 2005, Audit of signalised intersections in New Zealand: recommendations for practitioners, Road and Transport Research, vol. 14, no. 1, pp.72-77.

Austroads 2012 55

Effectiveness of Road Safety Engineering Treatments

C.7

Signal Visibility
Study Year 1991 Country US Environment/Treatment Replace pedestal with mast arms Reduction All crashes: 25% Angle: 63% Rear-end: 19% Left-turn: 35% increase All crashes: 28% Injury crashes: 17% Rear-end: 28% Angle: 35% PDO: 31% 35% All crashes: 10% All crashes: 7% Injury crashes: 3% PDO: 9%

Summary of research
Bhesania

Felipe, Mitic & Zein

1998

Possibly Canada

Additional primary head

Thomas & Smith Gan, Shen & Rodriguez Sayed, Leur & Pump

2001 2005 2005

US US US

Replace pedestal with mast arms 12-inch lens Improve visibility of signal heads (increase signal lens size, install new backboards, add reflective tape to existing backboards, and/or install additional signal heads) Yellow reflective tape added to backplate Add additional primary head

FHWA (includes references to a number of others already referenced)

2009

US

All crashes: 15% All crashes: 28% Angle: 35% Rear-end: 28% Fatal/injury: 17% PDO: 31% All crashes: 49 Fatal/injury: 44% PDO: 51% Right-turn: 12% Angle: 74% Rear-end: 41% All crashes: 7% Injury crashes: 3% PDO: 9% All crashes: 49 Fatal/injury: 44% PDO: 51% Right-turn: 12% Angle: 74% Rear-end: 41% All crashes: 3% Angle: 42% All crashes: 97 CMF Angle: 0.58 CMF 25% (95% CI -50; +5) 40% (95% CI -45; -35)

Pedestal to mast arm

Improve visibility Rodegerdts, Nevers & Robinson 2005 US Pedestal to mast arm

Srinivasan et al. Harkey et al. Elvik et al.

2008 2008 2009

US US International

12-inch lens 12-inch lens Secondary signal Improving sight conditions, signal heads & signal posts

Austroads 2012 56

Effectiveness of Road Safety Engineering Treatments

A number of studies provided figures for improving signal visibility. Three studies provided reductions for replacing a pedestal mount with a mast arm mount signal (Bhesania, Thomas & Smith, Rodegerdts, Nevers & Robinson). Based on these, a reduction of 35% is recommended (low confidence). Two studies provided reductions for the increasing lens size to 12 inches (Gan, Shen & Rodriguez and Srinivasan et al.). Based on these, a reduction of 6% (rounded to 5%) is recommended (low confidence). Two studies provided reductions for the provision of an additional signal head (Felipe, Mitic & Zein and Elvik et al.). Based on these, a reduction of 20% is recommended (medium confidence). Studies have also considered general improvements (Sayed, Leur & Pump) and addition of yellow reflective tape (FHWA). However, as there is generally only one study for each of these, a recommended crash reduction has not been provided. References assessed
Bahar, G, Masliah, M, Wolff, R & Park, P 2007, Desktop reference for crash reduction factors, report FHWASA-07-015, Federal Highway Administration, Washington DC, USA. Bhesania, RP 1991, Impact of mast-mounted signal heads on accident reduction, ITE Journal, vol.61, no.10, pp.25-9. Elvik, R, Hye, A, Vaa, T & Srensen, M 2009, The handbook of road safety measures, 2 edn, Emerald Publishing, Bingley, UK. Felipe, E, Mitic, D & Zein, SR 1998, Safety benefits of additional primary signal heads, Insurance Corporation Of British Columbia, Canada. FHWA Office of Safety 2009a, Traffic signals, Intersection safety issue briefs, issue brief 5, FHWA-SA-10005 Federal Highway Administration Office of Safety, Washington, DC. FHWA Office of Safety 2009b, Retroreflective borders on traffic signal backplate: a South Carolina success story, FHWA-SA-09-011, Federal Highway Administration Office of Safety, Washington, DC. Gan, A, Shen, J & Rodriguez, A 2005, Update of Florida crash reduction factors and countermeasures to improve the development of district safety improvement projects, Lehman Centre for Transportation Research, Miami, Florida. Harkey, DL, Srinivasan, R, Baek, J, Council, FM, Eccles, K, Lefler, N, Gross, F, Persaud, B, Lyon, C, Hauer, E & Bonneson, JA 2008, Accident Modification Factors for Traffic Engineering and ITS Improvements, NCHRP report 617, Transportation Research Board, Washington, DC, USA. Rodegerdts, LA, Nevers, B & Robinson, B 2005, Signalized intersections: informational guide, report FHWAHRT-04-091, Federal Highway Administration (FHWA), McLean, VA, USA. Sayed, T, Leur, P & Pump, J 2005, Safety impact of increased traffic signal backboards conspicuity Transportation Research Board annual meeting, 84th, 2005, Washington, DC, USA, TRB, Washington, DC, USA.
nd

Austroads 2012 57

Effectiveness of Road Safety Engineering Treatments

Srinivasan, R, Council, FM, Lyon, C, Gross, F, Lefler, NX & Persaud, BN 2008, Safety effectiveness of selected treatments at urban signalized intersections, Transportation Research Record, no. 2056, pp.70-6. Thomas, GB & Smith, DJ 2001, Effectiveness of roadway safety improvements, final report, Center for Transportation Research and Education, Iowa State University, USA.

Austroads 2012 58

Effectiveness of Road Safety Engineering Treatments

C.8

Channelisation at Intersections Splitter and Median Islands


Study Year 1984 1989 1991 1992 1997 1997 1997 1997 1997 2001 2001 2001 2001 2004 2004 2005 2005 2005 2005 Country Australia US Unknown UK NZ NZ NZ NZ NZ Australia Australia Australia UK US US US US US Australia Environment/Treatment Signalised intersections Unsignalised intersections Rural painted Splitter island on secondary road Rural painted Splitter all sites Rural intersections Urban intersections T-intersections X-intersections Traffic island on approach rural Traffic island on approach urban Channelisation Splitter islands Splitter islands (urban) Mountable median at intersection Non-mountable median at intersections Channelisation at intersections Mountable medians Non-mountable medians Traffic island on approach Seagull island Left-turn slip Indented right island 26% 54% 50% 10% 35% 44% 38% 45% 48% 43% 20% (not significant) 36% 36.4% 26.6% 29% 15% 25% All: 22% All: 15% Fatal: 90% All: 25% Pedestrian: 25% All: 18.7%,cas: 30.9% All: 5.7% increase, cas: 2.1% increase All: 11.1%, cas: 21.2% All: 49.6%, cas: 61.8% 30% Reduction

Summary of research
Teal in Ogden (1996) County Surveyors Society in Ogden (1996) Brude in Arndt (2004) Ward in Ogden (1996) LTSA LTSA LTSA LTSA LTSA BTE BTE Newstead & Corben Gorell & Tootill Shen et al. Shen et al. Gan, Shen & Rodriguez Gan, Shen & Rodriguez Gan, Shen & Rodriguez Meuleners et al.

Freeman et al.

2008

NZ and France (quoted other work) Various Various

Splitter island on minor approach T-intersection full channalisation (physical) T-intersection side road channalisation (physical)

Elvik et al. Elvik et al.

2009 2009

24% increase 18% increase

Austroads 2012 59

Effectiveness of Road Safety Engineering Treatments

Study Elvik et al.

Year 2009

Country Various

Environment/Treatment X-intersection full channalisation physical marked 32% 57% 20%

Reduction

Elvik et al.

2009

Various

X-intersection side road channalisation (physical)

C.8.1

General

Three studies provided general figures for channelisation at intersections (channelisation refers to situations where the island type is unclear; it may be a splitter, median island or both): Teal (1984), Gan, Shen and Rodriguez (2005), and Corben and Newstead (2001). The Teal (1984) figure has not been used, due to the age of the study. Therefore based on Gan et al. (2005) (22%), and Corben and Newstead (2001) (36.4%) a reduction of 29% may be assumed (rounded to 30%). There is low confidence in this figure. C.8.2 Splitter Islands

There were five studies that provided overall (general) reductions for the installation of splitter islands:

LTSA: 44% Corben and Newstead: 26.6% Brude: 10% Meuleners et al.: 30.9% Freeman et al.: 30%.

The average of these is 28% (rounded to 30%). There is medium confidence in this figure, as it is based on a number of studies, although there is quite a large variation in results from those studies. However, this reduction is in line with the general result for channelisation. When considering urban and rural environments:

Urban: Gorell and Tootill (2006), LTSA (1997) and BTE (2001) indicate reductions for urban environments (29%, 45% and 36% respectively). The average of these is 36% (rounded to 35%). There is medium confidence in this figure. Rural: LTSA (1997) and BTE (2001) also provided figures for rural intersections (38% and 20% respectively). As there are only two studies, and there is quite a large variation in the results, no conclusion is drawn. However, based on these two studies, it is likely that lower crash reductions would be achieved in rural environments than in urban environments.

Considering T-intersections versus cross-intersections, LTSA (1997) and Elvik et al. (2009) provided figures for splitter islands at T-intersections and cross-intersections. Due to mixed results, no conclusions are drawn. C.8.3 Median Islands

Only two studies considered median islands (islands on through roads at intersections). It was assumed that the benefit would be lower for median islands than for splitter islands, as other crash types not influenced by islands occur more regularly on through roads. It was not possible to
Austroads 2012 60

Effectiveness of Road Safety Engineering Treatments

distinguish between median and splitter islands from the BTE (2001) study, although this acted as a guide. The figures based on Shen et al. (2004) were therefore used as follows:

Install mountable median at intersection 15% reduction. Install non-mountable median at intersection 25% reduction.

There is low confidence in these figures as they are not based on any Australian or NZ research, and they are based on a single study in which the source of the figures is not known. References assessed
Arndt, O 2004, Relationship between unsignalised intersection geometry and accident rates, Doctor of Philosophy Thesis, Queensland University of Technology and Queensland Department of Main Roads, Brisbane, Qld. Bureau of Transport Economics (BTE) 2001, The black spot program 1996-2002: An evaluation of the first three years, report no. 104, Bureau of Transport Economics (BTE), Canberra, Australia. Elvik, R, Hye, A, Vaa, T & Srensen, M 2009, The handbook of road safety measures, 2 Publishing Group, Bingley, UK.
nd

edn, Emerald

Freeman, JR, Bansen, JA, Wemple, B & Spinks, R 2008, Innovative Operational Safety Improvements at Unsignalized Intersections, Kittelson & Associates for Florida Department of Transportation, Tallahassee, Florida, US. Gan, A, Shen, J & Rodriguez, A 2005, Update of Florida crash reduction factors and countermeasures to improve the development of district safety improvement projects, Lehman Centre for Transportation Research, Miami, Florida. Gorell, RSJ & Tootill, W 2001, Monitoring local authority road safety schemes using MOLASSES, report TRL 512, TRL, Crowthorne, UK. Land Transport Safety Authority 1997, Installation of throat & fishtail islands at intersections, Land Transport Safety Authority, New Zealand. Meuleners, L, Hendrie, D, Legge, M & Cercarelli, LR 2005, An evaluation of the effectiveness of the black spot programs in Western Australia, 2000-2002, The University of Western Australia, Crawley, WA. Newstead, S & Corben, B 2001, Evaluation of the 1992-1996 Transport Accident Commission funded accident black spot treatment program in Victoria, report 182, Monash University Accident Research Centre, Clayton, Vic. Ogden, KW 1996, Safer roads: a guide to road safety engineering, Avebury Technical, Aldershot, UK. Shen, J Rodriguez, A Gan, A & Brady, P 2004, Development and application of crash reduction factors: A state-of-the-practice survey of State Departments of Transportation, Transportation Research Board annual meeting, 83rd, Washington DC, TRB, Washington, DC, USA.

Austroads 2012 61

Effectiveness of Road Safety Engineering Treatments

C.9

Grade Separated Intersections


Study Year 1990 1990 1991 2004 Country Sweden Sweden UK Sweden Environment/Treatment Cross intersections T-intersections Not given Sinsen-Storo Oslo before and after study controlled for regression to the mean. Also compared results with those for Oslo in 1994 new interchange opened widened road from 4 to 6 lanes, moved tramline off road, grade separated (from signalised junction and roundabout). Speed went from 50 km/h (mostly) to 70 or 80 km/h. Minor access roads were also removed. T-junction instead of atgrade X-junction instead of at-grade Reduction 50% severity unknown 10% severity unknown 57% severity unknown Injury accident reductions of 51% (95% CI -68; -27) For Trondheim: 52% reduction (95% CI -71;-19) These figures take into account increase/decrease in traffic and regression to the mean

Summary of research
Hedman in Ogden (1996) Hedman in Ogden (1996) Walker & Lines in Ogden (1996) Amundsen & Elvik

Elvik et al.

2009

Europe

Unspecified severity: 16% reduction (95% CI -33; +4) Injury crashes: 24% reduction (95% CI -57 +33) Unspecified severity: 42% reduction (95% CI -52; -30) Injury crashes: 57% reduction (95% CI -62; -51) PDO: 36% reduction (95% CI -50; -19) Unspecified severity: 27% reduction (95% CI -36; -18) Injury crashes: 28% reduction (95% CI -40; -15) Unspecified severity: 15% reduction (95% CI -24; -5) Unspecified severity: 115% increase (95% CI +52; +205)

Elvik et al.

2009

Europe

Elvik et al.

2009

Europe

Signalised junction grade separated instead of at-grade Grade separated instead of partly at-grade Partly grade separated instead of at-grade X-junction with speed camera Partly grade separated instead of signalised junction Crossroad above instead of below main road Diamond instead of trumpet Diamond instead of junction with direct access ramps Diamond instead of clover-leaf

Elvik et al. Elvik et al.

2009 2009

Europe Europe

Elvik et al.

2009

Europe

Unspecified severity: 22% reduction (95% CI -41; +3) Unspecified severity: 4% reduction (95% CI -17; +10) Unspecified severity: 38% reduction (95% CI -59; -7) Unspecified severity: 25% reduction (95% CI -59; +40) Unspecified severity: 2% reduction (95% CI -19; +18)

Elvik et al. Elvik et al. Elvik et al.

2009 2009 2009

Europe Europe Europe

Elvik et al.

2009

Europe

Austroads 2012 62

Effectiveness of Road Safety Engineering Treatments

Study Elvik et al. Elvik et al. AASHTO

Year 2009 2009 2010

Country Europe Europe Unknown

Environment/Treatment Diamond instead of loop Diamond instead of other Convert at-grade to grade separated: Setting unspecified (four-leg intersection, traffic control unspecified) Traffic volume unspecified Convert at-grade to grade separated: Setting unspecified (three-leg intersection, traffic control unspecified) Traffic volume unspecified Convert at-grade to grade separated: Design = diamond, trumpet or cloverleaf with crossroad above freeway Setting unspecified Traffic volume unspecified

Reduction Unspecified severity: 9% reduction (95% CI -25; +10) Unspecified severity: 7% reduction (95% CI -17; +4) All severities CMF 0.58 (standard error 0.1) Injury CMF 0.43 (standard error 0.05) Non-injury CMF 0.64 (standard error 0.1)

AASHTO

2010

Unknown

All severities CMF 0.84 (standard error 0.2). Less reliable CMF due to standard error All severities CMF 0.73 (standard error 0.08) Injury CMF 0.72 (standard error 0.1)

AASHTO

2010

Unknown

All severities CMF 0.96 (standard error 0.1) Notation = observed variability suggests this treatment could result in an increase, decrease or no change in crashes

The crash reduction factor (CRF) for grade separation of cross intersections is 55%. This percentage is derived using the average of CRF data from Walker and Lines (57%), Hedman (50%), Elvik et al. (57%), Amundsen and Elvik (51% and 52%), and the AASHTO (57%). This CRF estimate is considered robust (i.e. high confidence) given the frequency of studies with similar CRF values. The CRF for T-intersections is 20%. This percentage is derived by taking the average of the Hedman (10%), Elvik et al. (24%) and AASHTO (28%) studies. This result is considered less reliable (medium confidence), as there are only three studies, and more variability in the results. References assessed
AASHTO 2010, Highway safety manual, 1 edn, American Association of State Highway and Transportation Officials, Washington, DC, USA. Amundsen, AH & Elvik, R 2004, Effects on road safety of new urban arterial roads, Accident Analysis & Prevention, vol. 36, no.1, pp.115-23. Elvik, R, Hoye, A, Vaa, T & Sorensen, M 2009, The handbook of road safety measures, 2 edn, Emerald Publishing Group, Bingley, UK. Ogden, KW 1996, Safer roads: a guide to road safety engineering, Avebury Technical, Aldershot, UK.
nd st

Austroads 2012 63

Effectiveness of Road Safety Engineering Treatments

C.10 Right-turn Lane Provision


Summary of research
Study Creasey and Agent Creasey and Agent Creasey and Agent Creasey and Agent Creasey and Agent Year 1985 1985 1985 1985 1985 Country US US US US US Environment/Treatment Non signalised right-turn with kerb Non-signalised right-turn painted Right-turn lane without signals Right-turn lane with signals Signalised right-turn channelisation (no right-turn phase) Right-turn lane Turn-lane construction Right-turn lane at signalised intersection Urban painted right-turn lane at intersection Right-turn lane painted or installed Four-leg rural unsignalised 60% 30% 25% 30% 15% Reduction

VicRoads Ermer, Fricker & Sinha Neuman in Ogden (1996) Harwood LTSA Harwood et al.

1990 1991 1993 1995 1994 2002

Australia US US US NZ US

40% 20% 1840% 35% 33% All crashes: 28% at intersections 55% on approach Casualty crashes: 35% at intersections 61% on approach

Harwood et al. Harwood et al.

2002 2002

US US

Four-leg rural signalised Four-leg rural newly signalised

All crashes: 18% All crashes: 35% at intersections 44% on approach Casualty crashes: 29% at intersections 42% on approach

Harwood et al.

2002

US

Four-leg urban unsignalised

All crashes: 27% at intersections 20% on approach Casualty crashes: 29% at intersections 55% on approach

Austroads 2012 64

Effectiveness of Road Safety Engineering Treatments

Study Harwood et al.

Year 2002

Country US

Environment/Treatment Four-leg urban signalised

Reduction All crashes: 10% at intersections 34% on approach Casualty crashes: 9% at intersections 35% on approach

Harwood et al.

2002

US

Four-leg urban newly signalised

All crashes: 24% at intersections 28% on approach Casualty crashes: 28% at intersections 43% on approach

Harwood et al.

2002

US

Three-leg rural unsignalised

All crashes: 44% at intersections 45% on approach Casualty crashes: 55% at intersections 44% on approach

Harwood et al. Harwood et al. Harwood et al. Harwood et al. Harwood et al. Harwood et al. Harwood et al. Harwood et al. Harwood et al. Harwood et al. Harwood et al. Harwood et al. Harwood et al. Agent, Stamatiadis & Jones

2002 2002 2002 2002 2002 2002 2002 2002 2002 2002 2000 2000 2000 1996

US US US US US US US US US US Various, but mainly US Various, but mainly US Various, but mainly US US

Three-leg rural signalised Three-leg urban newly signalised Three-leg urban unsignalised Three-leg urban signalised Three-leg rural unsignalised kerbed Three-leg rural unsignalised painted Four-leg urban signalised kerbed Four-leg urban signalised painted Four-leg rural unsignalised curbed Four-leg rural unsignalised painted Three-leg rural unsignalised Four-leg rural unsignalised Four-leg rural unsignalised Right-turn lane with signal

All crashes: 15% All crashes: 68% on approach All crashes: 33% at intersections 32% on approach All crashes: 7% 49% (all crashes) 43% (all crashes) 10% (all crashes) 9% (all crashes) 57% (all crashes) 23% (all crashes) 22% 24% (one approach on main road) 42% (both approaches on main road) 25%

Austroads 2012 65

Effectiveness of Road Safety Engineering Treatments

Study Agent, Stamatiadis & Jones LTSA BTE BTE Shen et al. Gluck et al. Gluck et al. Gluck et al. Gluck, Levinson & Stover Gluck, Levinson & Stover Gluck, Levinson & Stover Gluck, Levinson & Stover

Year 1996 1994 2001 2001 2004 1999 1999 1999 1999 1999 1999 1999

Country US NZ Australia Australia US US US US US US US US

Environment/Treatment Right-turn lane without signal Painted right-turn lanes Indented turn lane (left or right) urban Indented turn lane (left or right) rural Right-turn lane without signals Right-turn signalised Right-turn unsignalised Right-turn unsignalised physical separation Right-turn unsignalised painted separation Right-turn signalised physical separation Right-turn signalised painted separation Right-turn signalised (4-lane arterials) 35% 33% 32%

Reduction

28% (not signalised) 32% 41% 63% 65% 24% casualty 28.5% 40% 15% Rear-end: 17% increase and 59% reduction for entering vehicles due to greatly different results, these have not been averaged Angle: 19.5% (from 9 and 38) for entering vehicles Right-turn: 55% (from 56, 43, 54 and 66), study results for variety of scenarios: entering vehicles, per leg, and turning vehicles) Other: 51.5% (from 29 and 74) for entering vehicles

Gluck, Levinson & Stover

Right-turn unsignalised (4-lane arterials)

Rear-end: 79% (from 87, 62 and 88) for entering vehicles Angle: 17% increase (from 50% increase, 65% reduction and 68% increase) for entering vehicles Right-turn: 72.5% (from 37, 90, 77 and 86), study results for variety of scenarios: entering vehicles, per leg, and turning vehicles) Other: 49% (from 45 and 53) for entering vehicles

Arndt

2004

Australia

(Vogt 1999 study in four-leg intersections based in US) Channelised right-turn lane (unsignalised)

38%

Arndt

2004

Australia

Rear-end major: 98% lower than basic right-turn treatment Rear-end major: 97% lower than auxiliary right-turn treatment

Austroads 2012 66

Effectiveness of Road Safety Engineering Treatments

Study Arndt Meuleners et al. Gan, Shen & Rodriguez

Year 2004 2005 2005

Country Australia Australia US

Environment/Treatment Auxiliary Right-turn lane (unsignalised) Indented right island Right-turn lane (signalised intersection) Right-turn lane (unsignalised intersection) Painted lane (left/right not specified) Provide Right-turn lane

Reduction Rear-end major: 42% lower than basic right-turn treatment 49.6% All: 28% Right-turn: 40% Overturn: 28% All: 34% Right-turn: 55% All: 32% Rear-end: 75% All: 25% Rear-end: 65% Angle: 50% Side swipe: 20% Right-turn: 53% Injury: 17% All crashes: 27% All crashes: 19% Injury: 24% Injury: 4% Injury: 14% increase

Gan, Shen & Rodriguez Gan, Shen & Rodriguez Gan, Shen & Rodriguez

2005 2005 2005

US US US

Elvik et al.

2009

Various

Three-leg intersection: Physical/marked Physical Marked

Elvik et al.

2009

Various

X-intersection: Physical/marked Physical Marked

AASHTO

2010

US

Three-leg intersection: right-turn lane on 1 major approach: Rural unsignalised All crashes: 0.56 CMF Injury: 0.45 CMF All crashes: 0.67 CMF Injury: 0.65 CMF All crashes: 0.85 CMF All crashes: 0.93 CMF Injury: 0.94 CMF

Urban unsignalised

Rural signalised Urban signalised

Austroads 2012 67

Effectiveness of Road Safety Engineering Treatments

Study AASHTO

Year 2010

Country US

Environment/Treatment Four-leg intersection: right-turn lane on 1 major approach: Rural unsignalised

Reduction

All crashes: 0.72 CMF Injury: 0.65 CMF All crashes: 0.73 CMF Injury: 0.71 CMF All crashes: 0.82 CMF All crashes: 0.90 CMF Injury: 0.91 CMF All crashes: 0.76 CMF Injury: 0.72 CMF

Urban unsignalised

Rural signalised Urban signalised

Urban newly signalised AASHTO 2010 US Four-leg intersection: right-turn lane on 2 major approaches: Rural unsignalised

All crashes: 0.52 CMF Injury: 0.42 CMF All crashes: 0.53 CMF Injury: 0.50 CMF All crashes: 0.67 CMF All crashes: 0.81 CMF Injury: 0.83 CMF All crashes: 0.58 CMF Injury: 0.52 CMF

Urban unsignalised

Rural signalised Urban signalised

Urban newly signalised AASHTO 2010 US Four-leg intersection: right-turn lane on 4 approaches: Rural AASHTO 2010 US Four-leg intersection: right-turn lane on 2 major approaches: Rural AASHTO 2010 US Three-leg intersection: right-turn lane on major approaches: Rural

Injury: 0.73 CMF

Injury: 0.96 CMF

Injury: 0.73 CMF

Austroads 2012 68

Effectiveness of Road Safety Engineering Treatments

Study AASHTO

Year 2010

Country US

Environment/Treatment Three-leg intersection: right-turn lane on all approaches (major & minor): Rural Injury: 1.16 CMF

Reduction

C.10.1 All Locations Five studies gave an overall figure for the installation of right-turn lanes (Ermer, Fricker & Sinha, Meuleners et al., VicRoads, Gan, Shen & Rodriguez, and LTSA). These figures were 20% 49.6%, 40%, 25% and 33%. The average of these values is 34% (rounded to 35%, medium confidence). It is noted that Arndt (2004) reviewed the effectiveness of different types of right-turn lane installations (basic, auxiliary and channelised). Arndt found that channelised treatments experienced considerably lower crashes than basic and auxiliary treatments. C.10.2 Signals Four studies provided general reductions for installation of right-turn lanes at signalised intersections. Creasy and Agent 30%, Agent, Stamatiadis and Jones 25%, Gluck, Levinson and Stover 41% and Gan, Shen and Rodriguez 28%. Based on these, a reduction of 30% is recommended (medium confidence). Harwood et al. (2002) provided figures for three and four-leg intersections, as well as for intersection approach crashes. It is noted that Gluck, Levinson and Stover provided crash effectiveness according to a number of crash types (rear-end, angle, right-turn and other). An increase in rear-end type crashes was noted, with reductions in angle, right-turn and other crash types. While AASHTO provided some results for rural signalised intersections, these figures are not considered applicable in the Australian context. C.10.3 Unsignalised Intersections Five studies provided general reductions for installation of right-turn lanes at unsignalised intersections. Creasy and Agent 25%, Agent, Stamatiadis and Jones 35%, Shen et al. 32%, Gluck, Levinson and Stover 63% and Gan, Shen and Rodriguez 34%. The average of these figures was a 38% reduction, but given that the figure from Gluck et al. appeared to be an outlier, this was rounded down to 35% (medium confidence). It was noted that the benefit identified in a number of studies (e.g. Gluck, Levinson & Stover and Gan, Shen & Rodriguez) was generally greater for sites that were not signalised. The reductions derived for unsignalised and signalised is in line with the findings from Gluck, Levinson and Stover (1999) and Harwood et al. (2002), who also indicated that it appears that provision of a right-turn lane (painted or constructed) at unsignalised intersections may provide greater reductions than at signalised. It is noted that Gluck, Levinson and Stover provided crash reductions according to a number of crash types (rear-end, angle, right-turn and other).

Austroads 2012 69

Effectiveness of Road Safety Engineering Treatments

C.10.4 Rural Four studies identified reductions in rural areas. The BTE identified a 28% reduction in casualty crashes. For unsignalised T-intersections, Harwood et al. (2000) identified a 22% reduction, Harwood et al. (2002) identified a 44% reduction for all crashes (55% for casualty crashes), and AASHTO identified a 44% reduction for all crashes (55% for casualty crashes). The average of these figures (using casualty figures where available) was a 44% reduction, although a more conservative reduction of 40% is recommended (low confidence). For unsignalised X-intersections, Harwood et al. (2000) identified a 24% reduction (1 approach), Harwood et al. (2002) identified a 28% reduction for all crashes (35% for casualty crashes), and AASHTO (2010) identified a 28% reduction for all crashes (35% for casualty crashes). The average of these figures (using casualty figures where available) was a 31% reduction, rounded to 30% (medium confidence). Where reductions were available for 1 and 2 approaches, the reduction associated with the 1 approach treatment was selected, as this provides a more conservative estimate. Harwood et al. (2002) also provided reductions for intersection approach crashes (as well as the total intersection crashes). While AASHTO provided some results for rural signalised intersections, these figures are not considered applicable in the Australian and New Zealand context, as we tend not to have signals in rural environments. C.10.5 Urban Four studies identified reductions in urban areas. Harwood (1995) identified a 35% reduction of painted right-turn lanes at intersections in urban areas, while the BTE identified a 32% reduction for the provision of turn lanes, however this was for both right and left. Due to the limited information, a reduction has not been provided for this treatment type. For unsignalised T-intersections, Harwood et al. (2002) identified a 33% reduction for all crashes, and AASHTO identified a 33% reduction for all crashes (35% for casualty crashes). Based on these, a reduction of 35% is recommended. Given there are just two studies, there is a low level of confidence. For signalised T-intersections, Harwood et al. (2002) identified a 7% reduction for all crashes, and AASHTO identified a 7% reduction for all crashes (6% for casualty crashes). Based on these, a reduction of 5% is recommended. Given there are just two studies, and it is likely that these reductions are based on the same data, there is a low level of confidence. For unsignalised X-intersections, Harwood et al. (2002) identified a 27% reduction for all crashes (29% for casualty crashes), and AASHTO identified a 27% reduction for all crashes (29% for casualty crashes). Based on these, a reduction of 30% is suggested. Given there are just two studies, and it is likely that these reductions are based on the same data, there is a low level of confidence.

Austroads 2012 70

Effectiveness of Road Safety Engineering Treatments

For signalised X-intersections, Harwood et al. (2002) identified a 10% reduction for all crashes (9% for casualty crashes), and AASHTO (2010) identified a 10% reduction for all crashes (9% for casualty crashes). Based on these, a reduction of 10% is suggested. Given there are just two studies, and it is likely that these reductions are based on the same data, there is a low level of confidence. For newly signalised urban X-intersections, Harwood et al. (2002) identified a 24% reduction for all crashes (28% for casualty crashes). These reductions were also noted in AASHTO. Based on other evidence, it appears that reductions in rural areas were generally higher than in urban areas from the installation of right-turn lanes (see especially Harwood). Therefore, a figure of a 30% reduction in urban areas was selected as appropriate (low confidence). C.10.6 Painted and Physical A number of studies provided reductions for painted and/or physical right-turn lanes, summarised in the table below:
Study Creasy & Agent (1985) Ermer (1991) LTSA (1994) Harwood (1995) Gluck, Levinson & Stover. (1999) Harwood et al. (2002) Gan, Shen & Rodriguez (2005) Elvik et al. (2009) 33% 35% (Urban) Unsignalised 28.5%, signalised 15% Unsignalised three-leg rural 43%, four-leg rural Unsignalised 23% and signalised 9% 32% Three-leg 19% (all) and 14% increase for X-intersections (injury crashes) Three-leg 27% (all) and for X-intersections 4% (injury) Unsignalised 65% (24% casualty), signalised 40% Unsignalised three-leg rural 49%, four-leg rural Unsignalised 57% and urban signalised 10% 30% Painted 60% 20% Physical

Studies indicate that physical right-turn islands provide greater crash reductions than painted islands. Crash reductions for X-intersections appear to be higher than for T-intersection (Harwood et al. 2002 and Elvik et al.). For painted right-turn lanes, a mean of 32% was selected (based on the general figures from Creasy and Agent, LTSA, Harwood and Gan, Shen & Rodriguez), rounded to 30% (high confidence). Given this result, a reduction of 35% is recommended for the use of physical right-turn islands (low confidence). C.10.7 X and T-intersections A number of studies provided reductions for T and X-intersections, summarised in the table below:
Study Harwood et al. 2000 Harwood et al. 2002 22% 55% rural unsignalised (casualty) 15% rural signalised (all crashes) 33% urban unsignalised (casualty) 7% urban signalised 49% rural unsignalised kerbed 43% rural unsignalised painted T-intersection X-intersection 24% 1 approach, 42% two approaches 35% rural unsignalised (casualty) 18% rural signalised (all crashes) 29% urban unsignalised (casualty) (27% all) 9% urban signalised (10% all) 57% rural unsignalised kerbed 23% rural unsignalised painted

Austroads 2012 71

Effectiveness of Road Safety Engineering Treatments

Study Elvik et al. 2009 AASHTO 2010 17%

T-intersection 24% 27% (major approaches) 16% increase (all approaches) More effective unsignalised than X (1 approach)

X-intersection 4% (major approaches) 27% (all approaches) 2 approaches better than T More effective signalised than T (1 approach)

There is some information on the difference between installing right-turn lanes at X versus T-intersections, with a slight trend for a greater crash reduction benefit at unsignalised T-intersections, while for signalised intersections a greater benefit exists at X-intersections. However, given the lack of reliable information on this issue, no figure has been provided. This issue requires further research. References assessed
AASHTO 2010, Highway safety manual, 1st edn, American Association of State Highway and Transportation Officials, Washington, DC, USA. Agent, KR, Stamatiadis, N & Jones, S 1996, Development of accident reduction factors, research report KTC-96-13, Kentucky Transportation Centre, University of Kentucky, Lexington, KY, USA. Arndt, O 2004, Relationship between unsignalised intersection geometry and accident rates, Doctor of Philosophy Thesis, Queensland University of Technology and Queensland Department of Main Roads, Brisbane, Qld. Arndt, O & Troutbeck, R, 2004, Relationship between unsignalised intersection geometry and accident rates: final results, Road system and engineering forum 2004, Queensland Department of Main Roads, Brisbane, Qld, 32pp. Arndt, O & Troutbeck, R, 2006, New warrants for unsignalised intersection turn treatments, ARRB conference, 22nd, 2006, Canberra, ACT, Australia, ARRB Group, Vermont South, Vic, 19pp. Bureau of Transport Economics (BTE) 2001, The black spot program 1996-2002: An evaluation of the first three years, report no. 104, Bureau of Transport Economics (BTE), Canberra, Australia. Creasey, T & Agent, KR 1985, Development of accident reduction factors, research report UKTRP-85-6, Kentucky Transportation Research Program, University of Kentucky, Lexington, KY, USA. Elvik, R, Hye, A, Vaa, T & Srensen, M 2009, The handbook of road safety measures, 2 Publishing Group, Bingley, UK.
nd

edn, Emerald

Ermer, DJ, Fricker, JD & Sinha, KC 1991, Accident reduction factors for Indiana, JHRP-91-11, Purdue University, School of Civil Engineering, Lafayette, IN, USA. Gan, A, Shen, J & Rodriguez, A 2005, Update of Florida crash reduction factors and countermeasures to improve the development of district safety improvement projects, Lehman Centre for Transportation Research, Miami, Florida, USA. Gluck, J, Levinson, HS & Stover, V 1999, Impacts of access management techniques, NCHRP report 420, Transportation Research Board, Washington, DC. Harwood, DW 1995, Relationships between operational and safety considerations in geometric design improvements, Transportation Research Record, no.1512, pp.1-6.

Austroads 2012 72

Effectiveness of Road Safety Engineering Treatments

Harwood D.W., Council F.M., Hauer E., Hughes W.E. & Vogt A. 2000, Prediction of the expected safety performance of rural two-lane highways, report FHWA-RD-99-207, Federal Highway Administration (FHWA), McLean, VA, USA. Harwood, DW, Bauer, KM, Potts, IB, Torbic, DJ, Richard, KR, Kohlman Rabbani, ER, Hauer, E & Elefteriadou, L 2002, Safety effectiveness of intersection left- and right-turn lanes, report FHWA-RD02-89, Federal Highway Administration (FHWA), McLean, VA, USA. Land Transport Safety Authority 1994, Right-turn treatment, Land Transport Safety Authority, Wellington, New Zealand. Lyon, C, Persaud, B, Lefler, N, Carter, D & Eccles, K 2008, Safety evaluation of installing center two-way left-turn lanes on two-lane roads, Transportation Research Record, no. 2075, pp. 34-41. Meuleners, L, Hendrie, D, Legge, M & Cercarelli, LR 2005, An evaluation of the effectiveness of the black spot programs in Western Australia, 2000-2002, The University of Western Australia, Crawley, WA, Australia. Neuman TR, Pfefer R, Slack KL, Kennedy Hardy K, Harwood DW, Potts IB, Torbic DJ, Kohlman Rabbani ER 2003, Guidance for implementation of the AASHTO strategic highway safety plan: volume 5: A guide for addressing unsignalized intersection collisions, NCHRP report 500, Transportation Research Board, Washington, DC, USA. Ogden, KW 1996, Safer roads: a guide to road safety engineering, Avebury Technical, Aldershot, UK. Persaud, B, McGee, H, Lyon, C & Lord, D 2003, Development of a procedure for estimating expected safety effects of a contemplated traffic signal installation, Transportation Research Record, no. 1840, pp.96103. Shen, J Rodriguez, A Gan, A & Brady, P 2004, Development and application of crash reduction factors: A state-of-the-practice survey of State Departments of Transportation, Transportation Research Board annual meeting, 83rd, Washington DC, TRB, Washington, DC, USA. VicRoads 1990, Guidelines for the selection of projects under the road conditions sub-program (incorporating accident blackspot projects, mass action projects, railway level crossing projects), Road Safety Division, VicRoads, Kew, Vic. Yu, L, Qi, Y, Azimi, M, Gui, C & Guo, L 2007, Left-turn lane design and operation, for Texas Department of Transport, Department of Transport Studies, Texas Southern University, Houston, Texas, USA.

Austroads 2012 73

Effectiveness of Road Safety Engineering Treatments

C.11 Extend Right-turn Lane


Summary of research
Study Harwood et al. Harwood et al. Agent, Stamatiadis & Jones Neuman et al. Shen et al. Arndt Year 2002 2002 1996 2003 2004 2004 Country US US US US US Australia Unsignalised intersections Increase left-turn lane length (right for Australia) Environment/Treatment Urban signalised right-turn (NZ/Australia) Rural unsignalised right-turn (NZ/Australia) 29.7% increase 43% approach crashes 15% Reduction in rear-end crashes (extent unknown) 25% Rear-end-major vehicle accident rates on the short length right-turn slots were not significantly higher than that for the longer turn slots. This indicates that use of short turn slots may be an appropriate treatment to provide at sites with lower turning volumes 28% Significant decrease in rear-end crashes Reduction

Gan, Shen & Rodriguez Yu et al

2005 2007

US US

Increase turn-lane length (left/right not specified) Increase left-turn lane length (right for Australia)

The Harwood study showed an increase in crashes where right-turn lanes were extended at urban signalised intersections, and a 43% reduction at rural unsignalised intersections. While these were statistically significant results, they were based on very small sample sizes (3 and 4 respectively). Three other studies gave results for extending right-turn lanes (Agent, Stamatiadis & Jones 15%, Shen et al. 25% and Gan, Shen & Rodriguez 28%), all based on US research. Neuman et al. noted that extending turn lanes will reduce rear-end collisions, although the study did not provide an estimated reduction. They suggested that the reduction would likely be a function of a number of site characteristics including existing lane length, traffic volume and sight distance. However, Arndt (2004) observed that on major roads, rear-end crash rates were not significantly higher on the short turn lanes compared to the longer turn lane slots. Based on this it was decided not to provide an estimate of effectiveness. References assessed
Agent, KR, Stamatiadis, N & Jones, S 1996, Development of accident reduction factors, research report KTC-96-13, Kentucky Transportation Centre, University of Kentucky, Lexington, KY, USA. Arndt, O 2004, Relationship between unsignalised intersection geometry and accident rates, Doctor of Philosophy Thesis, Queensland University of Technology and Queensland Department of Main Roads, Brisbane, Qld. Arndt, O & Troutbeck, R, 2004, Relationship between unsignalised intersection geometry and accident rates: final results, Road system and engineering forum 2004, Queensland Department of Main Roads, Brisbane, Qld, 32pp.

Austroads 2012 74

Effectiveness of Road Safety Engineering Treatments

Gan, A, Shen, J & Rodriguez, A 2005, Update of Florida crash reduction factors and countermeasures to improve the development of district safety improvement projects, Lehman Centre for Transportation Research, Miami, Florida, USA. Harwood, DW, Bauer, KM, Potts, IB, Torbic, DJ, Richard, KR, Kohlman Rabbani, ER, Hauer, E & Elefteriadou, L 2002, Safety effectiveness of intersection left- and right-turn lanes, report FHWA-RD02-89, Federal Highway Administration (FHWA), McLean, VA, USA. Neuman TR, Pfefer R, Slack KL, Kennedy Hardy K, Harwood DW, Potts IB, Torbic DJ, Kohlman Rabbani ER 2003, Guidance for implementation of the AASHTO strategic highway safety plan: volume 5: A guide for addressing unsignalized intersection collisions, NCHRP report 500, Transportation Research Board, Washington, DC, USA. Shen, J Rodriguez, A Gan, A & Brady, P 2004, Development and application of crash reduction factors: A state-of-the-practice survey of State Departments of Transportation, Transportation Research Board annual meeting, 83rd, Washington DC, TRB, Washington, DC, USA. Yu, L, Qi, Y, Azimi, M, Gui, C & Guo, L 2007, Left-turn lane design and operation, for Texas Department of Transport, Department of Transport Studies, Texas Southern University, Houston, Texas, USA.

Austroads 2012 75

Effectiveness of Road Safety Engineering Treatments

C.12 Left-turn Lane Provision


Summary of research
Study Agent, Stamatiadis & Jones Harwood et al. Harwood et al. Harwood et al. Harwood et al. Harwood et al. Shen et al. BTE BTE Arndt Arndt Meuleners et al. Gan, Shen & Rodriguez Gan, Shen & Rodriguez Year 1996 2000 2002 2002 2002 2002 2004 2001 2001 2004 2004 2005 2005 2005 Country US US US US US US US Australia Australia Australia Australia Australia US US Environment/Treatment Left-turn lane Unsignalised three and four-leg Four-leg rural unsignalised Four-leg urban unsignalised Four-leg urban signalised Three-leg urban signalised Left-turn lane Indented right and left-turn urban Indented right and left-turn rural Left-turn lane Auxiliary left-turn lane (unslgnalised) Left-turn slip Painted lane Provide left-turn lane 25% One approach: 5% Both major approaches: 10% 14% at intersections 27% on approaches 40% at intersections 4% at signalised intersections 18% on approaches 45% on approaches 32% 32% 28% (not signalised) Channelised and auxiliary turn lanes similar safety benefits Rear-end major: 47% lower than basic left-turn treatment 11.1% All: 32% Rear-end: 75% All: 25% Rear-end: 65% Angle: 50% Side swipe: 20% Left-turn: 53% 28% Reduction

Gan, Shen & Rodriguez Elvik et al.

2005 2009

US Various

Increase turn land length (left/right not specified) Left-turn lane on: T or X-junction T-junction X-intersection

7% reduction 12% increase 19% reduction

AASHTO

2010

US

Left-turn lane on 1 major road approach: three or four-leg unsignalised three or four-leg signalised All crashes: 0.86 CMF Injury: 0.77 CMF All crashes: 0.96 CMF Injury: 0.91 CMF

Austroads 2012 76

Effectiveness of Road Safety Engineering Treatments

Study AASHTO

Year 2010

Country US

Environment/Treatment Left-turn lane on 2 major road approaches: three or four-leg unsignalised three or four-leg signalised

Reduction

All crashes: 0.74 CMF Injury: 0.59 CMF All crashes: 0.92 CMF Injury: 0.83 CMF

As a logic check, it was assumed that the benefit from a left-turn lane should be less than that for a right-turn lane (which produced an overall reduction of 34%). There were seven studies that provided an overall figure for the provision of a left-turn lane, as follows:

Agent, Stamatiadis and Jones: 25% Shen et al.: 32% BTE: 32% (urban) and 38% (rural) Meuleners et al.: 11.1% Gan, Shen and Rodriguez: 25% AASHTO: 23% (unsignalised) and 9% (signalised) Harwood et al. (2000): 5%.

The average of these is 22% (rounded to 20%). There is low confidence in this figure, as although it is based on a number of studies, there is quite a large variation in results (and therefore CRFs) from these studies. Where reductions were provided for the installation of a left-turn lane on one and two approaches (i.e. Harwood et al. 2000, AASHTO), the figure for one approach was used as this provides a more conservative estimate. The Harwood et al. (2002) and BTE studies provided reductions for rural and urban areas. BTE found a higher casualty crash reduction for urban areas (32%) than for rural (28%). Harwood et al. (2002) found mixed results with the highest crash reduction (all crashes) for urban unsignalised intersections (40%), followed by rural unsignalised (14%) and then urban signalised (4%). This study provided casualty crash reductions for rural unsignalised (23%) and urban signalised (9%). Note that the Harwood et al. results only applied to four-leg intersections. There are limited and mixed results regarding provision of left-turn lanes at T-intersections versus cross-intersections. Harwood et al. (2002) indicated a higher reduction for urban signalised three-leg intersections (45% on approach) compared to four-leg intersections (18% on approach), while Elvik et al. indicated a 12% increase for three-leg intersections compared to a 19% reduction for four-leg intersections. There are some indications that left-turns at rural T-intersections (at least in the US context) were associated with increased crash risk. It has been suggested that this might be a result of a higher proportion of turning movements rather than the provision of the turning lane itself. Gan, Shen and Rodriguez provided indicative crash reductions by crash type, however as only one study is available, it is not useful for this work.

Austroads 2012 77

Effectiveness of Road Safety Engineering Treatments

References assessed
AASHTO 2010, Highway safety manual, 1 edn, American Association of State Highway and Transportation Officials, Washington, DC, USA. Agent, KR, Stamatiadis, N & Jones, S 1996, Development of accident reduction factors, research report KTC-96-13, Kentucky Transportation Centre, University of Kentucky, Lexington, KY, USA. Arndt, O 2004, Relationship between unsignalised intersection geometry and accident rates, Doctor of Philosophy Thesis, Queensland University of Technology and Queensland Department of Main Roads, Brisbane, Qld. Arndt, O & Troutbeck, R, 2004, Relationship between unsignalised intersection geometry and accident rates: final results, Road system and engineering forum 2004, Queensland Department of Main Roads, Brisbane, Qld, 32pp. Arndt, O & Troutbeck, R, 2006, New warrants for unsignalised intersection turn treatments, ARRB Conference, 22nd, 2006, Canberra, ACT, Australia, ARRB Group, Vermont South, Vic, 19pp. Austroads 2009, Guide to road safety: part 8: treatment of crash locations, AGRS08/09, Austroads, Sydney, NSW. Bureau of Transport Economics (BTE) 2001, The black spot program 1996-2002: An evaluation of the first three years, report no. 104, Bureau of Transport Economics (BTE), Canberra, Australia. Elvik, R, Hye, A, Vaa, T & Srensen, M 2009, The handbook of road safety measures, 2 Publishing Group, Bingley, UK.
nd st

edn, Emerald

Gan, A, Shen, J & Rodriguez, A 2005, Update of Florida crash reduction factors and countermeasures to improve the development of district safety improvement projects, Lehman Centre for Transportation Research, Miami, Florida, USA. Harwood, DW, Council, FM, Hauer, E, Hughes, WE & Vogt, A 2000, Prediction of the expected safety performance of rural two-lane highways, report FHWA-RD-99-207, Federal Highway Administration (FHWA), McLean, VA, USA. Harwood, DW, Bauer, KM, Potts, IB, Torbic, DJ, Richard, KR, Kohlman Rabbani, ER, Hauer, E & Elefteriadou, L 2002, Safety effectiveness of intersection left- and right-turn lanes, report FHWA-RD02-89, Federal Highway Administration (FHWA), McLean, VA, USA. Meuleners, L, Hendrie, D, Legge, M & Cercarelli, LR 2005, An evaluation of the effectiveness of the black spot programs in Western Australia, 2000-2002, The University of Western Australia, Crawley, WA, Australia. Neuman, TR, Pfefer, R, Slack, KL, Kennedy, Hardy, K, Harwood, DW, Potts, IB, Torbic, DJ, Kohlman Rabbani ER 2003, Guidance for implementation of the AASHTO strategic highway safety plan volume 5: a guide for addressing unsignalized intersection collisions, NCHRP report 500, Transportation Research Board, Washington, DC, USA. Shen, J Rodriguez, A Gan, A & Brady, P 2004, Development and application of crash reduction factors: A state-of-the-practice survey of State Departments of Transportation, Transportation Research Board annual meeting, 83rd, Washington DC, TRB, Washington, DC, USA. Vogt, V & Bared, J 1998, Accident models for two-lane rural segments and intersections, Transportation Research Record, no. 1635, pp.18-29.

Austroads 2012 78

Effectiveness of Road Safety Engineering Treatments

C.13 Lane Width in Tunnels


Summary of research
Study Elvik et al. Year 2009 Country Unknown Environment/Treatment Unknown Reduction 40% reduction in injury crashes when increasing tunnel width from less than 6 m to more than 6 m.

No information was found on the effectiveness of changing lane widths in tunnels. Only one study was identified that appeared to be related to this issue, which found a 40% reduction for an increase in tunnel width. While this may be assumed to be linked with an increase in lane width, other factors may also vary with increasing tunnel width. Given the lack of guidance on this topic, additional information may be sought from literature on safety associated with lane widths generally. References assessed
Elvik, R, Hye, A, Vaa, T & Srensen, M 2009, The handbook of road safety measures, 2 Publishing Group, Bingley, UK.
nd

edn, Emerald

Austroads 2012 79

Effectiveness of Road Safety Engineering Treatments

C.14 Overtaking Lanes


Summary of research
Study Parker et al. Parker et al. Parker et al. Parker et al. Parker et al. Parker et al. Creasey & Agent Creasey & Agent Ermer, Fricker & Sinha Ermer, Fricker & Sinha Ermer, Fricker & Sinha VicRoads Ogden Slop & Catshoek McLean Thrush Agent, Stamatiadis & Jones Agent, Stamatiadis & Jones Agent, Stamatiadis & Jones Council & Stewart Koorey et al. Koorey et al. Koorey et al. Harwood et al. Harwood et al. Austroads Austroads Year 1983 1983 1983 1983 1983 1983 1985 1985 1991 1991 1991 1990 1992 1995 1996 1996 1996 Europe US NZ US USA USA USA USA USA Australia Rural Inter-urban roads USA Country Environment/Treatment Rural Rural Rural Reduction 25% reduction when passing lanes installed 38% less crashes where overtaking lanes are provided 2527% reduction in crashes 25% reduction 20% reduction from no passing zone 10% reduction from no passing sign Add passing lane results in 20% reduction Add climbing lane results in 10% reduction Construct passing blister = 20% reduction Construct travel lane = 10% reduction Upgrade no-passing zone = 30% reduction Install overtaking lane = 25% reduction No statistically significant difference with installation of overtaking lanes Prevention of all dangerous overtaking manoeuvres results in a 20% reduction 2538% reduction when overtaking lane constructed on 2 lane highways 38% less crashes in sites with overtaking lanes Recommend 20% reduction with installation of a passing or climbing lane, and 40% reduction in passing accidents with no passing zones Reported on mean reduction from literature of 48% for no passing zones, and 42% from state surveys Reported on mean reduction from lit of 28% from literature for use of passing / climbing lane, and 22% from State surveys Undivided 20% reduction when converting 2 lane to 4 lane (undivided) 13% reduction after construction of overtaking lane 54% reduction with full realignment 5% reduction with tack-on lane Passing lane = 25% reduction Short 4 lane section = 35% reduction Rural Sections with overtaking lanes have 1338% lower crash rates Sections with overtaking lanes have 25% lower crash rates. 2.5% reduction in advance (3 km) of overtaking lane, and for 5 km section following the overtaking lane Passing or climbing lanes in one direction Two-lane sections: 25% Four-lane sections: 35%

1996

US

1996

US

1999 1999 1999 1999 2000 2000 2001 2001

US NZ NZ NZ US US Australia Australia (NSW)

Gan, Shen & Rodriguez

2005

United States

Austroads 2012 80

Effectiveness of Road Safety Engineering Treatments

Study Elvik et al.

Year 2009

Country Various

Environment/Treatment

Reduction 13% reduction in injury for one direction only 40% reduction in injury crashes for both directions Report that benefit is greatest for higher volume roads

The mean of the reductions is 25%. There is a medium level of confidence in this figure, as although it is quite consistent across studies, it includes a range of different applications. Other figures indicate that the benefit of tack-on lanes is less than for new alignments and passing lanes (5% reduction versus a 54% reduction, although there is low confidence in these values). Benefits appear to be greater when passing lanes are used in two directions (i.e. four lanes with one through and one passing lane in each direction) than for one direction only (40% versus 13%). References assessed
AASHTO 2010, Highway safety manual, 1st edn, American Association of State Highway and Transportation Officials, Washington, DC, USA. Agent, KR, Stamatiadis, N & Jones, S 1996, Development of accident reduction factors, research report KTC-96-13, Kentucky Transportation Centre, University of Kentucky, Lexington, KY, USA. Austroads 2001, Economic evaluation of road investment proposals: improved prediction models for road crash savings, APR184/01, Austroads, Sydney, NSW. Austroads 2009, Guide to road safety: part 8: treatment of crash locations, AGRS08/09, Austroads, Sydney, NSW. Council, F & Stewart JR 1999, Safety effects of the conversion of rural two-lane to four-lane roadways based on cross-sectional models, Transportation Research Record, no. 1665, pp. 35-43. Creasey, T & Agent, KR 1985, Development of accident reduction factors, research report UKTRP-85-6, Kentucky Transportation Research Program, University of Kentucky, Lexington, KY, USA. Elvik, R, Hye, A, Vaa, T & Srensen, M 2009, The handbook of road safety measures, 2 Publishing Group, Bingley, UK.
nd

edn, Emerald

Ermer, DJ, Fricker, JD & Sinha,KC 1991, Accident reduction factors for Indiana, JHRP-91-11, Purdue University, School of Civil Engineering, Lafayette, IN, USA. Gan, A, Shen, J & Rodriguez, A 2005, Update of Florida crash reduction factors and countermeasures to improve the development of district safety improvement projects, Lehman Centre for Transportation Research, Miami, Florida, USA. Harwood, DW, Council, FM, Hauer, E, Hughes, WE & Vogt, A 2000, Prediction of the expected safety performance of rural two-lane highways, report FHWA-RD-99-207, Federal Highway Administration (FHWA), McLean, VA, USA. Hemion, RH 1969, A preliminary cost-benefit study of headlight glare reduction, report AR-683, Southwest Research Institute, San Antonio, TX, USA.

Austroads 2012 81

Effectiveness of Road Safety Engineering Treatments

Koorey, G, Farrelly, P, Mitchell, T & Nicholson, C 1999, Assessing passing opportunities: stage 2, research report 146, Transfund New Zealand, Wellington, NZ. Larsson, M, Candappa, N & Corben, B 2003, Flexible barrier systems along high speed roads: a lifesaving opportunity, report 210, Monash University Accident Research Centre, Clayton, Vic. McLean, J 1996, Review of accidents and rural cross section elements including roadsides, research report ARR 297, ARRB Transport Research, Vermont South, Vic. Mutabazi, MI, Russell, ER & Stokes, RW 1999, Review of the effectiveness, location, design, and safety of passing lanes in Kansas, report K-TRAN:KSU-97-1, Kansas Department of Transportation, Topeka, KS, USA, viewed 25 October 2010, <http://ntl.bts.gov/lib/7000/7500/7576/784476.pdf>. Ogden, KW 1992, Benefit/cost analysis of road trauma countermeasures: rural road and traffic engineering programs, report 34, Monash University Accident Research Centre, Clayton, Vic. Parker, MR, Flak, MA, Tsuchiyama, KH, Wadenstorer, SC & Hutcherson, F 1983, Geometric treatments for reducing passing accidents at rural intersections on two lane highways: volumes 1 and 2, FHWA/RD83-074/75, Office of Safety and Traffic Operations, Federal Highway Administration, Washington, DC, USA. Slop, M & Catshoek, JWD 1995, Recommended safety measures for application on interurban roads in the short term, report R-95-18, Institute for Road Safety Research, Leidschendam, Netherlands. Thrush, M 1996, Assessing passing opportunities: literature review, research report 60, Transit New Zealand, Wellington. VicRoads 1990, Guidelines for the selection of projects under the road conditions sub-program (incorporating accident blackspot projects, mass action projects, railway level crossing projects), Road Safety Division, VicRoads, Kew, Vic.

Austroads 2012 82

Effectiveness of Road Safety Engineering Treatments

C.15 Superelevation
Summary of research
Study Creasy and Agent Year 1985 Country United States Environment/Treatment Correcting/improving superelevation on all highway environments Correcting superelevation deviation on rural 2-lane rural highways Improving crossfall on a number of unknown road environments Modifying superelevation on all highway environments Correcting superelevation at intersections/ mid-blocks with high-speed traffic Superelevation deficiency (SD) correction for 2-lane rural highways 40% Reduction

Zegeer et al.

1992

United States

10%

Corben et al.

1996

United States

38% (not significant)

Agent, Stamatiadis & Jones Ogden

1996

United States

40%

1996

International

40% to 60% (head-on, overtaking, off-road on curve and out-of-control on curve)

Harwood et al.

2000

United States

Superelevation deficiency (SD) < 0.01 0.01 to < 0.02 0.02

AMF 1.00 (1.00 + 6(SD 0.01) (1.00 + 3(SD 0.02)

Hanley, Gibby & Ferrara

2000

United States

Superelevation modification and shoulder widening in unknown road environments Unknown road environment, relating to several treatments: superelevation correction, provide proper superelevation and resurfacing/superelevation Correcting superelevation deficiency for all types of crashes

Hanley (1996) study 12% to 13% (all crash types) Caltrans (1994) study 50% ITE (1992) study 75% 33% (run-off-road crashes) State survey 40% (all crash types) State survey 45% (all crash types) State survey 28% (all crash types)

Gan, Shen & Rodriguez

2005

United States

Gan, Shen & Rodriguez

2005

United States

Likely to be based on Harwood, as Harwood also includes this table, in addition to the CMFunction formulas Superelevation deficiency 0.02 0.03 0.04 0.05 AMF 1.06 1.09 1.12 1.15

Monsere et al.

2006

United States

Correcting superelevation deficiency for 2-lane rural highways

The following reductions are calculated using Harwood et al.s functions 15% (all crashes, correcting from 0.02 to 0.08) 11% (all crashes, correcting from 0.04 to 0.08) 6% (all crashes, correcting from 0.06 to 0.08)

Austroads 2012 83

Effectiveness of Road Safety Engineering Treatments

Study Austroads

Year 2009

Country Australia

Environment/Treatment Correcting superelevation in an unknown road environment Superelevation correction at blackspot sites

Reduction 50% (head-on crashes)

NSDOT

2009

United States

61% (all crash types)

Of the research reviewed, Harwood et al. (2000) and Zegeer et al. (1992) were considered to be the most comprehensive and robust. The Crash Modification Functions (CMFunctions) derived by Harwood et al (2000) were based on the work by Zeger et al (1992). Gan, Shen and Rodriguez (2005) and Monsere et al. (2006) both provided reductions based on the CMFunctions derived by Harwood et al (2000). Logically, the effect of improving superelevation deficiency would be a function, dependant on the original deficiency level. Therefore, these CMFunctions are recommended. Given that these functions are primarily based on one piece of work, and not based Australian or New Zealand findings there is a low level of confidence. Note that, superelevation, speed and curvature are all closely linked in terms of safety outcomes. Therefore a Crash Modification Function drawing these elements together would be more desirable. References assessed
Agent, KR, Stamatiadis, N & Jones, S 1996, Development of accident reduction factors, research report KTC-96-13, Kentucky Transportation Centre, University of Kentucky, Lexington, KY, USA. Austroads 2009, Guide to road safety: part 8: treatment of crash locations, AGRS08/09, Austroads, Sydney, NSW. Corben, B, Newstead, S, Diamantopoulou, K & Cameron, M 1996, Results of an evaluation of TAC funded accident black spot treatments, Combined 18th ARRB Transport Research conference and Transit New Zealand Land Transport symposium, 1996, Christchurch, New Zealand. ARRB Transport Research Ltd, Vermont South, Victoria, Australia. Creasey, T & Agent, KR 1985, Development of accident reduction factors, research report UKTRP-85-6, University of Kentucky, Kentucky Transportation Research Program, Lexington, KY. Gan, A, Shen, J & Rodriguez, A 2005, Update of Florida crash reduction factors and countermeasures to improve the development of district safety improvement projects, Lehman Centre for Transportation Research, Miami, Florida, USA. Hanley, K, Gibby, R. & Ferrara, T 2000, Analysis of accident-reduction factors on California State Highways, Transportation Research Board, Washington, DC, United States. Harwood, DW, Council, FM, Hauer, E, Hughes, WE & Vogt, A 2000, Prediction of the expected safety performance of rural two-lane highways, report FHWA-RD-99-207, Federal Highway Administration (FHWA), McLean, VA, USA. Illinois Department of Transport 2006, Crash reduction factors for HSIP program projects, Illinois Department of Transport, Illinois, USA. Monsere, C, Bertini, R, Breakstone, A, Bonner, C, Bosa, P, de la Houssaye, D, Horowitz, Z, HunterZaworski, K, 2006, Update and enhancement of ODOTs crash reduction factors, report no. FHWAOR-DR-06-11, Portland State University and Oregon State University, Oregon, USA. New York State Department of Transport 2009, PIES: reduction factor report, NYSDOT, Albany, New York.

Austroads 2012 84

Effectiveness of Road Safety Engineering Treatments

Ogden, KW 1996, Safer roads: a guide to road safety engineering, Avebury Technical, Aldershot, UK. Zegeer, C, Stewart, J, Council, F, Reinfurt, D, Hamilton, E 1992, Safety effects of geometric improvements on horizontal curves, Transportation Research Record, no.1356., pp.11-9.

Austroads 2012 85

Effectiveness of Road Safety Engineering Treatments

C.16 Impact Attenuators


Summary of research
Study Creasy and Agent Year 1985 Country United States Environment/Treatment Recommended CRFs for impact attenuators based on a range of speed and road environments Effectiveness of crash cushions in both rural and urban highway environment. CRF determined for interpolation of historical crash rates and comparing them to the observed crash rates after treatment Installation of impact attenuators in unknown highway environments. No further information Study of the effectiveness of impact attenuators in unknown highway environments Reduction 60% (fatal crashes) 10% (injury crashes)

Griffin

1986

United States

78% (fatal crashes) 27% (injury crashes)

Ogden

1996

United States

75% (fatal and serious injury crashes)

Agent, Stamatiadis & Jones

1996

United States

State survey Average 29% (all crashes) Average 75% (fatal crashes) Average 50% (injury crashes) Review of literature Average 31% (all crashes) Average 65% (fatal crashes) Average 36% (injury crashes)

Agent et al. in Bahar et al. (2007)

1996

United States

Study of the effectiveness of impact attenuators in unknown highway environments Study of the effectiveness of impact attenuators in unknown highway environments Installation of impact attenuators on a range of highway environments across the United States Study of the effectiveness of impact attenuators in a range of speed and road environments

Average 5% (all crash types) Average 75% (fatal crashes) Average 50% (injury crashes) Average 29% (all severities) Average 82% (fatal crashes) Average 50% (injury crashes) Average 29% (all crashes) Average 83% (fatal crashes)

Gan et al. in Bahar et al. (2007)

2005

United States

Gan, Shen & Rodriguez

2005

United States

FHWA

2008

United States

29% for all crash types (all crash severities) 75% for all crash types (fatal crashes) 50% for all crash types (injury crashes) 45% reduction in run-off-road crashes (all crash severities) 69% in run-off-road into fixed objects (fatal crashes) 69% in run-off-road into fixed objects (injury crashes) 46% in run-off-road into fixed objects (PDO crashes) 41% (all crashes) 45% (run-off-road crashes)

ADOT

2009

United States

Study of the effectiveness of impact attenuators in unknown highway environments

Austroads 2012 86

Effectiveness of Road Safety Engineering Treatments

Study Elvik et al.

Year 2009

Country Various

Environment/Treatment Installation of impact attenuators

Reduction 69% (fatal crashes, 95% level of significance, CI: -83, -46) 69% (injury crashes, 95% level of significance, Cl: -75, -62) 46% (PDO, 95% level of significance, CI: -63, -23) 69% (fatal crashes, standard error 0.3) 69% (injury crashes, standard error 0.1) 46% (PDO crashes, standard error 0.3)

AASHTO

2010

United States

Describes the effectiveness of various impact attenuators on: rural two-lane roads, rural multilane highways, freeways, expressways and urban/suburban arterials

It should be noted that provision of impact attenuators may not reduce the incidence of crashes, but rather influences the severity outcomes when a crash does occur (i.e. injury crash types are likely to transfer to non-injury crashes). The average casualty crash reduction was determined to be 40%. However, due to changes in vehicle and barrier technology, this figure was deemed to be too conservative and some of the older studies such as Creasey and Agent (1985) and Griffin (1983) were excluded. Therefore, a 50% crash reduction was chosen, and this was considered to have a medium level of confidence. For fatal crashes, the average CRF amongst all the studies was 72% (rounded to 70%) and this figure was deemed to have high confidence. References assessed
AASHTO 2010, Highway safety manual, 1 edn, American Association of State Highway and Transportation Officials, Washington, DC, USA. ADOT 2009, Benefit/cost ratio economic analysis, section 231, in Traffic engineering policies, guides and procedures (PGP), Arizona Department of Transportation, Phoenix, AZ. viewed 25 July 2011, < http://www.azdot.gov/highways/Traffic/standards/PGP/TM231.pdf>. Agent, KR, Stamatiadis, N & Jones, S 1996, Development of accident reduction factors, research report KTC-96-13, Kentucky Transportation Centre, University of Kentucky, Lexington, KY, USA. Bahar, G, Masliah, M, Wolff, R & Park, P 2007, Desktop reference for crash reduction factors, report FHWASA-07-015, Federal Highway Administration, Washington DC, USA. Creasey, T & Agent, KR 1985, Development of accident reduction factors, research report UKTRP-85-6, University of Kentucky, Kentucky Transportation Research Program, Lexington, KY. Elvik, R, Hye, A, Vaa, T & Srensen, M 2009, The handbook of road safety measures, 2 Publishing Group, Bingley, UK.
nd st

edn, Emerald

Federal Highway Administration 2008, Toolbox of countermeasures and their potential effectiveness for roadway departure crashes, report no. FHWA-SA-07-013, Federal Highway Administration, Washington DC, viewed 22 July 2011, <http://safety.fhwa.dot.gov/tools/crf/resources/briefs/rdwydepartissue.pdf>.

Austroads 2012 87

Effectiveness of Road Safety Engineering Treatments

Gan A, Shen J & Rodriguez A 2005, Update of Florida Crash Reduction Factors and Countermeasures to improve the Development of District Safety Improvement Projects, Lehman Centre for Transportation Research, Miami, Florida, viewed 25 March 2008. Griffin, L 1983, How effective are attenuation devices (crash cushions) in reducing deaths and injuries? Texas Transportation Researcher, vol. 19, no. 3, pp. 6-7. Ogden, KW 1996, Safer roads: a guide to road safety engineering, Avebury Technical, Aldershot, UK.

Austroads 2012 88

Effectiveness of Road Safety Engineering Treatments

C.17 Signs Advisory


Summary of research C.17.1 Curve Warning Signs
Study Creasey & Agent Sanderson et. in Donald Year 1985 1985 Country USA Not known, quoted in Australian study Australia Environment/Treatment Curves Two-way highways Reduction 30% reduction all crashes 30% reduction from advance warning signs

Moses

1987

Introduction of small advance direction/warning signs Warning signs curve warning

17% reduction in rear-end crashes 15% reduction in total crashes 37% average reduction all crashes (lit review 11 papers) 32% average reduction all crashes (state survey 16 states) 30% reduction run-off-road crashes (from recommended reduction factors table) 30% reduction in head-on crashes

Agent, Stamatiadis and Jones

1996

USA

Federal Office of Road Safety, in Austroads 2004 Elvik et al.

1996

Australia

Non-intersection

2009

International

Unknown

Injury accident reduction of 30% (95% CI -73;+84) No study had significant results Install new fluorescent curve signs or upgrade existing to fluorescent rural All severities, non-intersection CRF 18% All severities, non-intersection, head on, run-off-road, side swipe CRF 18% Fatal/serious injury/minor injury, non-intersection CRF 25% All severities, night, non-intersection CRF 35% All severities, night, non-intersection, head-on, run off-road, side swipe CRF 34%

Srinivasan et al.

2009

USA

A mean crash reduction factor of 25% was derived by taking the average of Elvik et al. (30%), Creasy and Agent (30%), Moses (15%), Sanderson et al. (30%), Srinivasan et al. (18%) and Agent, Stamatiadis and Jones (37%). Given the consistency between the studies, a high level of confidence can be placed in this CRF. Additional crash type specific CRFs were given by Moses (17% for rear-end crashes), Agent, Stamatiadis and Jones (30% for run-off-road crashes) and FORS (30% for head-on crashes).

Austroads 2012 89

Effectiveness of Road Safety Engineering Treatments

C.17.2 Speed Advisory Signs


Study Kneebone Year 1964 Country Australia Environment/Treatment NSW Reduction Speed advisory signs 1st study 62% reduction in casualty crashes, 56% reduction in all crashes (Hume Highway) had bad crash history Speed advisory signs 2nd study highly significant 70% reduction in number of casualties, 25% reduction in all crashes Speed advisory signs 3rd study: 21% reduction in all crashes Warning signs advisory speed 30% average reduction all crashes (lit review two papers) 26% average reduction all crashes (state survey two states) Injury accidents reduced by 13% (95% CI -22 -2) results based on three studies from 1972 or earlier Reduction factor is based on no signage vs. speed advisory sign CMF is 0.87 reduction with 0.09 standard error Note two of the studies used here are a double-up with Elvik

Agent, Stamatiadis and Jones

1996

USA

Not specified

Elvik et al.

2009

International

Unknown

AASHTO

2010

USA

Unknown

An average CRF of 40% was derived from Elvik et al. (13%), Kneebone (62% casualties, 70% casualties, 21% all crashes) and Agent, Stamatiadis and Jones (30%). Due to the inconsistency of results there is a low level of confidence in this figure. References assessed
AASHTO 2010, Highway safety manual, 1st edn, American Association of State Highway and Transportation Officials, Washington, DC, USA. Agent, KR, Stamatiadis, N & Jones, S 1996, Development of accident reduction factors, report KTC-96-13, Kentucky Transportation Centre, University of Kentucky, Lexington, KY, USA. Andreassen, DC 1989, Strategies for safety problems, research report ARR 163, Australian Road Research Board, Vermont South, Vic. Andrew OBrien and Associates 2000, Managing traffic flow on urban freeways: appendix A: literature review, Andrew OBrien and Associates, Melbourne, Vic. Austroads 2009, Guide to road safety: part 8: treatment of crash locations, AGRS08/09, Austroads, Sydney, NSW. Bureau of Transport Economics (BTE) 2001, The black spot program 1996-2002: An evaluation of the first three years, report no. 104, Bureau of Transport Economics (BTE), Canberra, Australia. Creasey, T & Agent, KR 1985, Development of accident reduction factors, research report UKTRP-85-6, Kentucky Transportation Research Program, University of Kentucky, Lexington, KY, USA.

Austroads 2012 90

Effectiveness of Road Safety Engineering Treatments

Donald, D 1997, Be warned! A review of curve warning signs and curve advisory speeds, research report 304, ARRB Transport Research, Vermont South, Vic. Elvik, R, Hye, A, Vaa, T & Srensen, M 2009, The handbook of road safety measures, 2 Publishing Group, Bingley, UK.
nd

edn, Emerald

Ewing, R 1999, Traffic calming: state of the practice, report FHWA-RD-99-135, Federal Highway Administration, Washington, DC, USA. Kneebone, DC 1964, Advisory speed signs and their effect on traffic, Australian Road Research Board conference, 2nd, 1964, Melbourne, Victoria, Australian Road Research Board, Vermont South, Vic, vol. 2, no. 1, pp. 524-41. Kulmala, R 1994, Measuring the safety effect of road measures at junctions, Accident Analysis and Prevention, vol. 26, no. 6, pp. 781-94. Lamm, R, Zumkeller, K & Beck, A 2001, Traffic safety: the relative effectiveness of a variety of road markings and traffic control devices, Road Safety on Three Continents, 2000, Pretoria, South Africa, VTI Konferens 15A, Swedish National Road and Transport Research Institute, Linkoeping, Sweden, pp. 120-32. Montella, A 2009, Safety evaluation of curve delineation improvements: empirical Bayes observational before-and-after study, Transportation Research Record, no. 2103, Transportation Research Board, Washington, pp. 6979. Moses, P 1987, Combating the road toll, National Local Government Engineering conference, 4th, 1987, Perth, Western Australia, Institution of Engineers Australia, Canberra, ACT, pp.70-4. Srinivasan, R, Baek, J, Carter, D, Persaud, B, Lyon, C, Eccles, K, Gross, F, & Lefler, N 2009, Safety evaluation of improved curve delineation, report FHWA-HRT-09-045, Federal Highway Administration, Washington, DCUK-MoRSE 2010, UK-MoRSE, Greensafe Foundation, Birmingham, UK, viewed 15 December 2010, <www.uk-morse.com>. UK-MoRSE 2010, UK-MoRSE, Greensafe Foundation, Birmingham, UK, viewed 15 December 2010, <www.uk-morse.com>. Winnett, MA & Wheeler, AH 2002, Vehicle-activated signs: a large scale evaluation, report 548, TRL, Crowthorne, UK.

Austroads 2012 91

Effectiveness of Road Safety Engineering Treatments

C.18 Pedestrian Treatments


Summary of research C.18.1 Pedestrian Fencing and Barriers
Study Teale Year 1984 Country Australia Environment/Treatment Safety bars (Pedestrian fencing to separate vehicles and pedestrians) Reduction 14% reduction in crashes (90% confidence interval 4% to 24%) 20% reduction in pedestrian crashes for fences which obscure the drivers view of pedestrians 48% reduction in pedestrian crashes for fences that obstructed the motorists view to a lesser extent UK before and after with control. Number of locations not known Ordinary fences obscuring drivers view of pedestrians associated with 20% reduction in pedestrian crashes Fences obscuring drivers view to lesser extent associated with a 48% reduction in pedestrian crashes. Was especially good for children as they are shorter Unknown Installed barrier fences along 18 sections of road in 1969. Pedestrian crossing crashes reduced by nearly 20% 29% reduction (95% CI -52 to -5) in pedestrian crashes

Stewart

1988

England

Retting, Ferguson & McCartt

2003

UK

Campbell et al.

2004

Japan

Elvik et al.

2009

International

Unknown

Information sources used were Teale (14% reduction), Stewart (20% for low visible fencing), Campbell et al. (20%) and Elvik et al. (29%). Retting, Ferguson and McCartt was excluded because it reports the result from Stewart. The average CRF based on these figures was 21% (rounded to 20%) with a medium level of confidence. The CRF is likely to be higher if fencing that does not obscure visibility to pedestrians is used. C.18.2 Improved Lighting
Study Pegrum, in Campbell et al. Retting, Ferguson & McCartt Year 1972 2003 Country Australia Israel Environment/Treatment Urban Urban Reduction 62% decrease in night-time pedestrian crashes Night-time pedestrian crashes decreased by 57%. Also non-significant decrease in daytime pedestrian crashes of 21%. At control sites there was a non-significant decrease in night-time pedestrian crashes of 60% 60% reduction in vehicle hits pedestrian crashes Improved lighting at pedestrian crossings. Based on 2 studies that used a control group. Neither controlled for exposure CRF 63% reduction in pedestrian accidents at night (95% CI -79;-36)

Austroads Elvik et al.

2004 2009

Australia International

Non-intersection Unknown

Austroads 2012 92

Effectiveness of Road Safety Engineering Treatments

Relevant information sources used: Elvik et al. 63%, Pegrum 62%, Austroads 60%, Retting, Ferguson and McCartt 57%. Average CRF based on these four sources is 60%. Medium confidence is placed in this CRF. All the CRFs are closely aligned and the average CRF is close to Elviks figure. C.18.3 Marked Crossings
Study Teale Teale Moses Zeeger et al. cited in Campbell et al. Koespell et al. in Retting, Ferguson & McCartt 2003 Year 1984 1984 1987 2001 Country Australia Australia USA USA Environment/Treatment Zebra crossing Zebra crossing with channelisation Not specified Not specified Reduction 66% increase in crashes (90% CI -175%; +43%) 10% reduction (90% CI -97%; +100%) Ratio of pedestrian crashes for marked and unmarked crossings was 6:1 On 2 lane roads, marked crosswalk was associated with no difference in pedestrian crash rate Crash risk was 2.1 times greater at sites with a marked crosswalk, almost all of the excess risk was due to 3.6 fold higher risk associated with marked crosswalks at sites with no traffic signal or stop sign Marked crosswalk with in-pavement lights alerting to pedestrian presence Israel 4 urban uncontrolled pedestrian crossings. Before and after with no controls. Rate of vehicle pedestrian conflict reduced to <1%, from 1% to 7% USA one urban uncontrolled crossing. Before and after with no controls. No CRF given percentage of drivers not yielding to pedestrians dropped from 31% to 8% 40% reduction in vehicle hits pedestrian crashes Adding or improving a crosswalk was associated with a CRF of 25% in three states Remove marked unprotected crosswalks from arterial intersections. One study. CRF of 73% Elvik et al. only used studies that controlled for pedestrian and vehicle volume with the following results: Pedestrian crashes on 2 lane roads Pedestrian crashes on multi-lane roads Pedestrian crashes on all roads AASHTO 2010 2 lane and multilane roads with AADT <12 000 UK CRF 8% (95% CI -43+51). Non-significant CRF +88% (95% CI -32+424). Stat significant increase CRF +44% (95% CI -6+121) Marked vs. unmarked crosswalk has no statistically significant effect on pedestrian crashes 24% decrease (95% CI -36.65; 83.99)

2002

USA

Six cities

Retting, Ferguson & McCartt

2003

Israel, USA

Unknown

Austroads Shen et al. Bahar et al. Elvik et al.

2004 2004 2007 2009

Australia USA Unknown

Non-intersection Unknown Urban

UK-MoRSE website

2010

Austroads 2012 93

Effectiveness of Road Safety Engineering Treatments

The CRF falls into three categories for marked crossings: 2 lane roads, multi-lane roads and the overall CRF. Studies of impact when used on 2 lane roads were Elvik et al. (8% reduction, but not significant), AASHTO (no effect) and Zegeer et al. (no change). The multi-lane road study found an 88% increase (Elvik et al.) and a statistically significant increase (Zegeer et al.) but no effect if AADT was under 12 000 (AASHTO). The overall studies were performed by Elvik et al. (44% increase), Shen et al. (25% decrease) Moses (increase), Zegeer (increase), Koespell (increase), Austroads (40% decrease in pedestrian crashes), Teale (66% increase, although not significant), Bahar et al. (10% decrease with channelization), UK-MoRSE (17% decrease) and Retting, Ferguson and McCartt (increase). None of the studies provided information about whether exposure effects had been factored into their calculations (particularly issues relating to increased pedestrian exposure at the location where crossings are installed). There is too little data to come to a conclusion about two-lane roads, and marked crossings are not used on multilane roads in Australia. There is a lot of variation in the overall studies but these indicate that there may be an increase in pedestrian crashes when marked crossings are installed. However, this increase may be due to increased exposure resulting from pedestrians choosing to use the crossing rather than cross nearby. C.18.4 Pedestrian Overpasses
Study Creasey & Agent VicRoads Retting, Ferguson & McCartt Year 1985 1990 2003 Country USA Australia Japan Environment/Treatment Not specified Not specified Urban Reduction 95% reduction in pedestrian crashes from construction of pedestrian crossover 10% reduction in total crashes from an overpass Pedestrian vehicle crashes decreased by 91% within 100 m of the overpasses and 85% within 200 m, crashes unrelated to pedestrians crossing the road increased 14% within 100 m of the overpasses and 23% within 200 m 90% reduction in vehicle hits pedestrian crashes from an overpass CRF ranged from 67% to 100% (data from four states) Shen typically only based on surveys of states Reviewed 3 studies CRF ranged from 5% to 100% based on Shen et al. (2004) Reduction of 82%

Austroads Shen et al.

2004 2004

Australia USA

Non-intersection Unknown

Bahar et al.

2007

Reviewed US studies International meta-analysis

Unknown

Elvik et al.

2009

Unknown

Relevant information sources used: Creasey and Agent 95%, VicRoads 10% of all crash types (not just pedestrian), Austroads 90%, Bahar et al. 90%, Elvik et al. 82% and Retting, Ferguson and McCartt 91%. Average CRF 90%. As a low level of confidence can be placed in this CRF due to the work, other than Elvik et al. lacking detail this CRF is rounded down to 85%.

Austroads 2012 94

Effectiveness of Road Safety Engineering Treatments

C.18.5 Pedestrian Signals


Study Teale Creasey & Agent Austin, Martin & Fox Year 1984 1985 1996 Country Australia USA Sussex Unknown Two sites with signalised pedestrian crossings Urban intersections USA 1297 intersections in 15 cities Sweden one intersection in small town; one intersection in Stockholm Environment/Treatment Reduction New pedestrian signals: 39% reduction (90% CI +32, + 46) After adding a pedestrian signal phase there was a 60% reduction in pedestrian crashes Signal timing was altered for shorter delay times 44% for one site and 24% for the other 37% reduction in pedestrian and bicycle crashes Two studies on exclusive pedestrian signal phase USA comparative analysis showed risk of pedestrian/vehicle crashes halved at exclusive timing intersections when compared with standard pedestrian signals Sweden small town intersection had 24% reduction in conflicts. Stockholm intersection had 10% reduction (not significant) Reduction of pedestrian crashes of 70% USA USA and Canada Unknown Unknown Unknown Going from no crosswalk to signalised pedestrian crossing: Mid-block Intersection Elvik et al. 2009 Unknown Going from marked crosswalk to signalised pedestrian crossing Add exclusive pedestrian phase one US study: 34% reduction range was 7% to 60% 0% to 55% (two studies)

Retting in Bahar et al. (2007) Retting, Ferguson & McCartt

2002 2003

USA USA Sweden

Austroads Bahar et al. Bahar et al. Elvik et al.

2004 2007 2007 2009

49% (not significant, CI -81 +35) 2% (CI -48 +84. Most studies did not use comparison group 27% reduction (confidence interval -59 +29).

Pedestrian signal treatments fell into four categories: improve/change signal timing, exclusive pedestrian phase, new signal installation and changing from a marked to a signalised crossing. Austin, Martin and Fox found CRFs of 24% and 44% with improved signal timing whereas Retting (cited in Bahar et al.) found a 37% reduction by changing signal timing (note includes pedestrians and cyclists). When a pedestrian phase was added Bahar et al. found a 34% reduction (range 7 to 60%), Creasy and Agent a 60% reduction, Austroads a 70% reduction and Retting, Ferguson and McCartt a 50% reduction. Installation of new signals resulted in a 39% reduction in pedestrian crashes (Teale), a non-significant 49% reduction (Elvik et al. mid-block) and a 2% reduction (Elvik et al. intersection). Changing a marked crossing into a signalised crossing resulted in a 27% reduction (Elvik et al.). Improved signal timing: CRF 35%, low confidence as this is based on few studies (one of which gave a combined CRF for pedestrians and cyclists.

Austroads 2012 95

Effectiveness of Road Safety Engineering Treatments

Pedestrian phase: CRF 50% (the average was 53.5% and was rounded down); low confidence due to the large range in Bahar et al. study and the source of the Austroads data is unclear. New signals: CRF cannot be calculated as there are too few studies at this time. Additionally the three included are not able to be compared as Teales installation point is unknown and the other two are installed at mid-blocks and intersections and different effects may be experienced depending on the installation point. Marked crossing to signalised: CRF cannot be calculated as there is only one study. C.18.6 Refuges
Study Teale Moses Garder, in Retting, Fergison and McCartt VicRoads LTSA Year 1984 1987 1989 1990 1995 Country Australia Australia Sweden Australia New Zealand Refuge replacing zebra crossing Urban intersections Not specified Not specified Environment/Treatment 80.7% Risk of pedestrian-vehicle conflicts decreased by roughly two thirds 30% reduction in total crashes from refuges/channelisation/kerb extension 33% reduction in pedestrian crashes where either or both pedestrian refuges and bulbous kerbs (kerb extension) were used 18% reduction for sites with only pedestrian refuges 37% reduction for sites with only bulbous kerbs 57% crash reduction Pedestrian facilities/refuge 5.8% reduction in casualty crashes (not significant) Pedestrian facilities/refuge 30.1% reduction in casualty crashes (not significant) 50% to 60% 50% reduction in vehicle hits pedestrian crashes 25% to 69% 43% reduction 34% reduction refuge, repainted markings, from 2 lanes to one lane Reduction 38% reduction (confidence interval +28 +48)

Ewing BTE BTE Retting, Ferguson and McCartt Austroads Bahar et al. Elvik et al. Doherty, Poole and Mintz-Roth

1999 2001 2001 2003 2004 2007 2009 2010

USA Australia Australia Sweden, USA Australia Various Worldwide USA/New York city

Not specified, summary of numerous studies Urban Regional Unknown Non-intersection Unknown Worldwide Metropolitan

A total of nine studies were considered when deciding on the CRF for pedestrian refuges: 50%, 60%, 38%, 43%, 18%, 30%, 57%, 50%, 66%. The average CRF was 45.7%. A further refinement was trialled where the less robust study CRFs were removed. These were 38%, 30% and 50%. This gave a CRF of 49%. It was decided to adopt the CRF of 45% as a more conservative figure as many of the studies were not from Australasia.

Austroads 2012 96

Effectiveness of Road Safety Engineering Treatments

C.18.7 Roundabouts
Study Brilon et al. in Retting, Ferguson and McCartt (2003) Schoon & van Minnen in Retting, Ferguson and McCartt (2003) Midson Bahar et al. Year 1993 Country Germany Environment/Treatment Intersections converted from traffic signals or stop signs Intersections converted from traffic signals or stop signs Tasmanian shopping strip Unsignalised intersection to roundabout Reduction 75% reduction in pedestrian crashes

1994

Netherlands

73% reduction in pedestrian vehicle crashes

2009 2007

Australia Flanders

75% reduction in pedestrian crashes 27%

A CRF of 60% was selected based on the average of Midson (75% reduction), Bahar et al. (27% reduction), Brilon et al. (75% reduction) and Schoon and van Minnen (73% reduction). There is low confidence in this CRF given the variance, and only one study being from Australia. Note, as this treatments relates to the provision of roundabouts (and not a specific pedestrian treatment) this has been included under Intersection treatments roundabouts in Table 3.1. C.18.8 Raised Pedestrian/Wombat Crossings
Study Geoplan 1994 in Austroads 2000 Bahar et al. Elvik et al. Year 2000 2007 2009 Country Australia USA International Environment/Treatment Not specified Unknown Unknown Reduction 8% reduction in pedestrian crashes 8% reduction in pedestrian crashes and 30% reduction in all crashes When a raised crosswalk goes in where no crosswalk previously existed: 65% reduction in all accidents (no information on pedestrian reduction) When a raised crosswalk goes in where a marked crosswalk had been a 42% (95% CI -70 to +11) reduction in pedestrian crashes

Elvik et al.

2009

International

Unknown

There is a mean CRF of 20% for pedestrian crashes based on the average of Bahar et al. (8%), Geoplan (8%) and Elvik et al. (42%). Crashes of all types could be expected to reduce from this measure. There is low confidence in this figure given the range of results. References assessed
AASHTO 2010, Highway safety manual, 1st edn, American Association of State Highway and Transportation Officials Washington, DC, USA. Archer, J, Candappa, N, & Corben, B, 2008, Effectiveness of the dwell-on-red signal treatment to improve pedestrian safety during high-alcohol hours, Australasian road safety research policing and education conference, 2008, Adelaide, South Australia, Monash University Accident Research Centre, Clayton, Vic, pp. 232-45. Austin, K, Martin, B & Fox, H 1996, Improving pedestrian priority in UTC systems, European transport th forum, 24 , seminar H, Brunel University, United Kingdom, PTRC Education and Research Services Ltd, London, UK, vol. P407, 12pp. Austroads 2000, Pedestrian and cyclist safety: recent developments, report AP-R155/00, Austroads, Sydney, NSW.

Austroads 2012 97

Effectiveness of Road Safety Engineering Treatments

Austroads 2009, Guide to road safety: part 8: treatment of crash locations, AGRS08/09, Austroads, Sydney, NSW. Bahar, G, Masliah, M, Wolff, R & Park, P 2007, Desktop reference for crash reduction factors, report FHWASA-07-015, Federal Highway Administration, Washington DC, USA. Brindle, R & Morrisy, Z 1998, Local area traffic management: review and survey of effectiveness, ARRB contract report CR RC6047-1, ARRB Transport Research, Vermont South, Vic. Bureau of Transport Economics (BTE) 2001, The black spot program 1996-2002: An evaluation of the first three years, report no. 104, Bureau of Transport Economics (BTE), Canberra, Australia. Campbell, BJ, Zegeer, CV, Huang, HH & Cynecki, MJ 2004, A review of pedestrian safety research in the United States and abroad, report FHWA-RD-03-042, Federal Highway Administration Virginia, USA, viewed on 26 April 2004, <http://www.walkinginfo.org/pdf/PedSynth/Ped_Synthesis_Report.pdf>. Candappa, N, Fotheringham, N, Lenn, M, Corben, B, Johansson, C & Smith, P 2005, Evaluation of an alternative pedestrian treatment at a roundabout, Australasian road safety research policing education conference, Wellington, New Zealand, Ministry of Transport, Wellington, NZ, 10 pp. Catchpole, J & Cairney, PT 1990, Accidents at major/minor intersections: a challenge for the traffic engineering profession, Local Government Engineers Association of Western Australia state th conference, 7 , Perth, Western Australia, Local Government Engineers Association of Western Australia, Perth, WA, technical papers, vol. 2, 12 pp. Cleaver, S, Jurisich, I & Dunn, R 2007, Safety implications of flush medians in Auckland City: further analyses, report 312, Land Transport New Zealand Research, Wellington, NZ, 86pp. Corben, BF & Cunningham, JA 1989, Traffic engineering treatment of hazardous locations, in KW Ogden & th DW Bennett (eds), Traffic engineering practice, 4 edn, Monash University, Clayton, Vic, pp.271-88. Corben, BF, Ambrose, C & Chee Wai, F 1990, Evaluation of accident black spot treatments, report 11, Monash University Accident Research Centre, Clayton, Vic. Creasey, T & Agent, KR 1985, Development of accident reduction factors, research report UKTRP-85-6, University of Kentucky, Kentucky Transportation Research Program, Lexington, USA. Davies, H & Winnett, M 1993, Why do pedestrian accidents happen? European summer annual meeting, st 21 , traffic management and road safety, 13-17 September, University of Manchester Institute of Science and Technology, PTRC, London, vol. P 365, pp.31524. Doherty, AM, Poole, H & Mintz-Roth, J 2010, Designing safe streets for seniors in New York City, ITE st Technical conference and exhibit: Meeting transportations 21 century challenges, 15-17 March 2010, Savannah Georgia, USA, ITE, Philadelphia, Pennsylvania, USA, 10pp. Elvik, R, Hye, A, Vaa, T & Srensen, M 2009, The handbook of road safety measures, 2 Publishing Group, Bingley, UK.
nd

edn, Emerald

Ewing, R 1999, Traffic calming: state of the practice, report FHWA-RD-99-135, Federal Highway Administration, Washington, DC, USA. Gorell, RSJ & Tootill, W 2001, Monitoring local authority road safety schemes using MOLASSES, report TRL 512, TRL, Crowthorne, UK.

Austroads 2012 98

Effectiveness of Road Safety Engineering Treatments

Hoareau, E, Newstead, S & Cameron, M 2006, An evaluation of the default 50 km/h speed limit in Victoria, report 261, Monash University Accident Research Centre, Clayton, Vic. Imberger, K, Cairney, P, Boschert, L & Styles, T 2004, Pedestrian countermeasure research project, contract report RC4019-2, ARRB Transport Research, Vermont South, Vic. Jordan, P 1995, Road safety audit: what it can do to improve safety for pedestrians, Australian pedestrian and bicyclist safety and travel workshop, 1994, Melbourne, Victoria, Australia, ARRB Transport Research, Vermont South, Vic, pp.213-22. Jurisich, I, Segedin, T, Dunn, R & Smith, M 2003, Experience of using flush medians in Auckland City, 26 Australasian Transport Research Forum, 1-3 October 2003, Wellington, New Zealand, NZ Institute of Highway Technology, Plymouth, NZ, 21pp. Kennedy, J & Sexton, B 2009, Literature review of road safety at traffic signals and signalised crossings, report PPR436, TRL, Crowthorne, UK, viewed 6 September 2010,
<http://www.trl.co.uk/online_store/reports_publications/trl_reports/cat_road_user_safety/report_literature_review_ of_road_safety_at_traffic_signals_and_signalised_crossings.htm>.
th

Kumar, A 1990, Road safety benefits from skid resistance program, report for Road Safety Division Roads Corporation Victoria, Sinclair Knight & Partners, South Melbourne, Vic. Land Transport Safety Authority 1995, Accident investigation monitoring analysis version 3.0, Land Transport Safety Authority, Wellington, NZ. Lenn, MG, Corben, BF & Stephan, K 2007, Traffic signal phasing at intersections to improve safety for alcohol-affected pedestrians, Accident Analysis and Prevention, vol.39, no.4, pp.751-6. Maxwell, A, Kennedy, J & Routledge, I 2010, Study compares accident frequency at Puffins and crossings using farside facilities, Traffic Engineering and Control, vol.51, no.8, pp.317-21. Markowitz, F, Sciortino, S, Fleck, JL & Bond PE 2006, Pedestrian countdown signals: experience with an extensive pilot installation, ITE Journal, vol.76, no.1, pp. 43-8. McLean, AJ & Anderson, RWG 2008, Metrication of the urban speed limit and pedestrian fatalities Australasian road safety research policing education conference, 2008, Adelaide, South Australia, Department for Transport, Energy and Infrastructure, SA, 7pp. Michael, LG, Bruce, CF, & Karen, S 2007, Traffic signal phasing at intersections to improve safety for alcohol-affected pedestrians, Accident Analysis & Prevention, vol. 39, no.4, pp.751-56. Midson, K 2009, Traffic control measures to improve pedestrian safety in shopping strips, Public Works Engineering, April/May, pp.40-6. Moses, P 1987, Combating the road toll, National local government engineering conference, 4th, 1987, Perth, WA, Institution of Engineers, Canberra, ACT, pp.70-4. Retting, RA, Ferguson, SA & McCartt, AT 2003, A review of evidence-based traffic engineering measures designed to reduce pedestrian-motor vehicle crashes, American Journal of Public Health, vol.93, no.9, pp.145663. Shen, J Rodriguez, A Gan, A & Brady, P 2004, Development and application of crash reduction factors: A state-of-the-practice survey of State Departments of Transportation, Transportation Research Board annual meeting, 83rd, Washington DC, TRB, Washington, DC, USA.

Austroads 2012 99

Effectiveness of Road Safety Engineering Treatments

Stewart, D 1988, Pedestrian guardrails and accidents, Traffic Engineering and Control, vol. 29,no. 9, pp. 450-5. Teale, G 1984, The evaluation of the effectiveness of low cost traffic engineering projects, consultant report CR 22, Office of Road Safety, Canberra, ACT. UK-MoRSE 2010, UK-MoRSE, Greensafe Foundation, Birmingham, UK, viewed 15 December 2010, <www.uk-morse.com>. Van Houten, JE, Malenfant, JE, Van Houten, J & Retting, R 1997, Using auditory pedestrian signals to reduce pedestrian and vehicle conflicts, Transportation Research Record, no. 1578, pp.20-22. VicRoads 1990, Guidelines for the selection of projects under the road conditions sub-program (incorporating accident blackspot projects, mass action projects, railway level crossing projects), Road Safety Division, VicRoads, Kew, Vic. Webster, N 2006, The effect of newly installed Puffin crossings on collisions, London Road Safety Unit Research Report, Transport for London, London, UK. Zegeer, CV, Blomberg, R, Henderson, D, Masten, S, Marchetti, L, Levy, MM, Fan, Y, Sandt, L, Brown, A, Stutts, J, & Thomas, L 2008, Evaluation of Miami-Dade pedestrian safety demonstration project, Transportation Research Record, no. 2073, , pp.1-10.

Austroads 2012 100

Effectiveness of Road Safety Engineering Treatments

C.19 Transverse Rumble Strips


Summary of research
Study Harwood Gunatillake Elvik et al. Srinivasan, Baek & Council Year 1993 2001 2009 2010 Country United States Australia/NZ Various Unspecified Environment/Treatment Installation of transverse rumble strips Installation of transverse rumble strips Rumble strips in front of junctions Installation of transverse rumble strips at intersections Rumble strips on at stop controlled approach Rumble strips on approach to intersections Curves Curves Railway crossing approach Minor approach to Tintersections Railway crossing approach (stop controlled crossings) On intersection approach (Stop controlled) Reduction 14100% (Harwood expressed concerns regarding validity) 2050% (Unspecified crash type) 33% injury crashes 21% (Fatal and injury crashes) 31% (Fatal and serious injury crashes) 24% increase (PDO crashes) (All at 90% confidence interval) 28% (all crashes) 90% rear-end Speeds tend to reduce in most cases, but not to a practical level 4.8 km/h reduction in speed 4 km/h reduction in speed 4.8 km/h reduction in speed 5 km/h reduction in speed

Gan, Shen & Rodriguez Gan, Shen & Rodriguez McGee & Hanscom Barker cited in Charman et al. (2010) Hore-Lacy Hore-Lacy Radalj and Kidd

2005 2005 2006 2007 2008 2008 2005

United States United States United States United Kingdom Victoria Victoria Australia

Thompson, Burris & Carlson

2006

United States

1.6 km/h reduction in speed

Five studies were found that identified crash reductions associated with the installation of transverse rumble strips. In addition to these, many studies concentrated on speed reduction associated with this treatment. Two of the five studies identified a wide range in treatment effectiveness. Srinivasan, Baek and Council (2010) found that at sites with rumble strips there was a significant decrease in fatal and injury crashes (21%) and an even higher decrease in fatal and serious injury crashes (31%). However, there was a significant rise in property damage only crashes (24%), all at the 90% confidence level. Elvik et al. identified a 33% reduction in injury crashes in advance of junctions. Gan, Shen and Rodriguez (2005) identified a 28% reduction for all crashes where rumble strips were provided on the approach to stop controlled intersections and a reduction of 90% was determined for rear-end crashes. Based on these three studies the average reduction is 27% for all crashes (rounded to 25%, low confidence).

Austroads 2012 101

Effectiveness of Road Safety Engineering Treatments

References assessed
Charman, S, Grayson, G, Helman, S, Kennedy, J, de Smidt, O, Lawton, B, Nossek, G, Wiesauer, L, Frds, A, Pelikan, V, Skldan, P, Pokorn, P, Matejka, M, & Tucka, P2010, Self-explaining roads literature review and treatment information, deliverable Nr 1, Road ERA net, Europe. Elvik, R, Hye, A, Vaa, T & Srensen, M 2009, The handbook of road safety measures, 2 edn, Emerald Publishing, Bingley, UK. Gan A, Shen J & Rodriguez A 2005, Update of Florida Crash Reduction Factors and Countermeasures to improve the Development of District Safety Improvement Projects, Lehman Centre for Transportation Research, Miami, Florida. Gunatillake, T 2001, Investigation of the use and design of rumble strips, contract report: RC1877-1, ARRB Transport Research Ltd, Vermont South, Vic. Hore-Lacy, 2008, Rumble strip effectiveness at rural intersections and railway level crossings contract report VC73896-1, ARRB Group, Vermont South, Australia. Harwood, DW 1993, Use of rumble strips to enhance safety, National Cooperative Highway Research Program synthesis of highway practice 191, Transportation Research Board, Washington, DC. McGee, H & Hanscom, F 2006, Low-cost treatments for horizontal curve safety, report no. FHWA-SA-07-002 Federal Highway Administration Washington DC. Radalj, T & Kidd, B 2005, A trial with rumble strips as a means of alerting drivers to hazards at approaches to passively protected railway level crossings on high speed Western Australian rural roads Australasian road safety research policing education conference, Wellington, New Zealand, Ministry of Transport, Wellington, NZ, 11pp. Srinivasan, R, Baek, J & Council, F2010, 'Safety evaluation of transverse rumble strips on approaches to stop-controlled intersections in rural areas', Journal of Transportation Safety & Security, vol. 2, no.3, pp. 261-78. Thompson, TD, Burris, MW & Carlson, PJ 2006, Speed changes due to transverse rumble strips on approaches to high-speed stop-controlled intersections, Transportation Research Record, no. 1973, pp. 1-9.
nd

Austroads 2012 102

Effectiveness of Road Safety Engineering Treatments

C.20 Street Closure


Summary of research
Study Brownfield in Brindle (1984) Year 1980 Country UK Environment/Treatment LATM measures, principally street closures Reduction Internal roads Total crashes: 31.2% reduction (non-significant) Pedestrian crashes: 24% reduction (non-significant) Peripheral roads Overall reduction (non-significant) (extent of reduction not listed) Reduction from 43 to 2 crashes (95%) Reduction from 41 to 18 crashes (56%) Injury crashes: 64% reduction

Bagby in Brindle (1984) Smith & Appleyard in Brindle (1984) Pfundt in Brindle (1984)

1980 1980 1980

United States, Grand Rapids United States, California Germany

Street diverters and closures LATM applications Closures, constrictions and meandering pavements with alternate angled parking LATM programs elimination of through traffic

Department of Transport in Brindle (1984)

1978

Australia

Internal roads Woodville South: 48% Flinders Park: 33% reduction Peripheral roads Crash rate reported as unchanged Internal roads Injury crashes: 56% reduction Total crashes: 42% reduction Peripheral roads Injury crashes: 32% reduction Internal roads Reduction from 50 to 29 crashes (42%) Peripheral roads Between major intersections - reduction from 145 to 127 crashes (12%) The major intersection rate remained about the same 50% to 80% reduction in vehicle crashes 10% to 50% reduction in vehicle and pedestrian crashes Reducing the number of median crossings and intersections on urban and suburban arterials, appears to reduce the number of intersection and driveway-related crashes. However, the magnitude of the crash effect is not certain at this time

Cairney & Brebner in Brindle (1984)

1980

Australia

Street closures South Australia (Unley)

Cairney & Brebner in Brindle (1984)

1980

Australia

Street closures South Australia (Burnside)

Ogden

1996

United States

Street closure at low-speed crossintersection and link closure on a mid-block Access management at intersections

AASHTO

2010

USA

No recent information about the effect of street closures was found. Brindle (1984) reviewed a number of LATM treatment studies that included street closures. A number of studies (Brownfield, Department of Transport and Cairney & Brebner) distinguished between crash effects for internal and peripheral roads. Internal roads generally refer to situations where the street closure occurs within the treated area, while peripheral roads are at the perimeter of the treated area. Five studies provided crash reductions for internal roads, with an average of 42% (rounded to 40%). There is a medium level of confidence in this figure. For peripheral roads, there was a range in the reported effects from little or no change up to a 32% reduction. A crash reduction of 5% may be assumed for peripheral roads (with a low level of confidence).

Austroads 2012 103

Effectiveness of Road Safety Engineering Treatments

References assessed
AASHTO 2010, Highway safety manual, 1 edn, American Association of State Highway and Transportation Officials, Washington, DC, USA. Brindle, RE 1984, Town planning and road safety: a review of literature and practice, special report 28, Australian Road Research Board, Vermont South, Vic. Ogden, KW 1996, Safer roads: a guide to road safety engineering, Avebury Technical, Aldershot, UK.
st

Austroads 2012 104

Effectiveness of Road Safety Engineering Treatments

C.21 Street Lighting


Summary of research
Study Thorpe in Andreassen (1989) Wortman & Lipinski in Isebrands et al. Roberts & Walker in Isebrands et al. Hall & Fisher Carstens & Berns in Isebrands et al. Creasey & Agent Year 1962 1972 1976 1977 1984 1985 USA/Iowa US based research (Kentucky) Country Australia USA/Illinois USA/Iowa Environment/Treatment Very good street lighting only Rural Rural All Rural General 20% (significant) 30% night crashes 49% night crashes 30% No significant reduction detected 50% Reduction

Intersections Routes Interchanges Railway crossing Osram-Gec Box in Isebrands et al. Ermer, Fricker & Sinha 1985 1987 1991 Australia USA/Illinois USA/Indiana All Urban intersection New installation Luminaire replacement New bridge lighting LTSA 1995 NZ All Intersections Routes New installations Agent, Stamatiadis & Jones 1996 US based research (Kentucky) General

55% 50% 50% 60% 30% 21% night crashes (significant) 37% 25% upgrade 16% 59% 33% 41% 30% 38% 33% lighting upgraded 50% night crashes

Road segments Intersections Interchanges Train crossings Corben et al. LTSA 1997 1997 Australia NZ NZ NZ NZ Preston & Schoenecker 1999 USA/Minnesota Rural and metro Upgraded lighting Intersections upgraded lighting Routes upgraded lighting New installations Rural intersections

45% night crashes 50% night crashes 50% night crashes 60% night crashes 24.9% increase in casualty crashes 33% reduction night crashes (significant) 41% reduction night crashes (significant) 30% reduction night crashes (significant) 38% decrease in night 40% reduction night crashes (significant) 15% reduction all crashes (significant)

Austroads 2012 105

Effectiveness of Road Safety Engineering Treatments

Study Bruneau, Morin & Pouliot

Year 2001

Country Canada

Environment/Treatment Rural four-lane motorway Continuous lighting (straight sections and interchanges) Rural four-lane Interchanges Urban Rural Urban Rural Rural intersections

Reduction 33% reduction night (significant) for injury and PDO

Canada BTE Gorell & Tootill Isebrands et al. Shen et al. 2001 2001 2004 2004 Australia UK USA/Minnesota USA

Interchange 4% reduction (not significant) 27% increase (significant) in all (night and day) 63.2% reduction (significant) 31% decrease in all 8% reduction in all 35% reduction night crashes (significant) Daytime crashes increased, but not significant For all crashes no injury/PDO breakdown General illumination CRF average of three states 27% (range 1937%) Night-time reduction, general illumination CRF average of two states 40% (range 3050%) For all crashes no injury/PDO breakdown Improved street lighting CRF two states both had reductions of 25% Night-time reduction, improved street lighting CRF two states both had reductions of 50%

Shen et al.

2004

USA

Shen et al.

2004

USA

Mid-blocks

For all crashes no injury/PDO breakdown Install/improved lighting at roadway segment average of three states 23% (range 2025%) Night-time reduction, install/improve lighting at roadway segment average five states 36% (range 2045%) For all crashes no injury/PDO breakdown Install/improve lighting at intersections average of three states CRF 32% (range 3036%) Night-time reduction, install/improve lighting at intersections average of six states CRF 55% (range 5067%) For all crashes no injury/PDO breakdown Install/improve lighting at interchanges two states both had CRF 25% Night-time reduction, install/improve lighting at interchanges three states had a CRF of 50% For all crashes no injury/PDO breakdown Average of three states CRF 41% (range 3062%) Night-time reduction, install/improve lighting at railway crossings three states each had CRF of 60% Pedestrians fatal 78% CRF (standard error 87) Pedestrians injury 42% CRF (standard error 18) All fatal/injury crashes CRF of 17% Night-time but all crash severities CRF of 50%

Shen et al.

2004

USA

Intersections

Shen et al.

2004

USA

Intersections

Shen et al.

2004

USA

Install/improve lighting at railway crossings

Bahar et al.

2007

Majority seem to be US based

Intersections improve intersection lighting

Bahar et al.

2007

Majority seem to be US based

Install lighting

Austroads 2012 106

Effectiveness of Road Safety Engineering Treatments

Study Bahar et al.

Year 2007

Country Majority seem to be US based Majority seem to be US based Majority seem to be US based Majority seem to be US based Majority seem to be US based International

Environment/Treatment Mid-blocks improve lighting Mid-blocks improve lighting Mid-blocks improve lighting Mid-blocks improve lighting Install lighting at interchanges CRF for going from unlit to lit

Reduction Fatal rural highway CRF of 73% (standard error 72) Based on a meta-analysis Fatal urban highway CRF of 63% (standard error 52) Based on a meta-analysis Injury crashes rural highway CRF of 20% (standard error 12) Based on a meta-analysis Injury urban highway CRF of 31% (standard error 7) Based on a meta-analysis Fatal/injury CRF of 26% (standard error 38) Night-time crashes all road types fatal CRF 60% (95% CI -62 -57) (controlled for publication bias CRF is 14% and not controlled for publication bias CRF is 23%; the source did not report 95% CI for these) Severity unknown: head-on crashes CRF 52% (95% CI -57 -46) Head-on crashes CRF 20% (95% CI -54; +44) Severity unknown: rear-end crashes CRF 54% (95% CI -68 -33) Rear-end crashes CRF 41% (95% CI -71; +21) Severity unknown: single vehicle crashes CRF 39% (95% CI -64 +3) Single vehicle crashes CRF 5% (95% CI -50; +79) Night-time crashes rural areas Fatal CRF 87% (95% CI -98; -34) (controlled for publication bias CRF is 14% and not controlled for publication bias CRF is 26%; the source did not report 95% CI for these) Night-time crashes at intersections/junctions rural areas Injury CRF 22% (95% CI -28; -15) Night-time crashes urban areas Fatal CRF 43% (95% CI -61; -15) Injury CRF 29% (95% CI -34; -23) Night-time pedestrian crashes urban areas Fatal CRF 78% (95% CI -88; -62) Injury CRF 50% (95% CI -57; -43) Night-time crashes at intersections/junctions urban areas Injury CRF 40% (95% CI -51; -27) Night-time crashes on motorways (controlled for publication bias CRF is 4% and not controlled for publication bias CRF is 13%; the source did not report 95% CI for these)

Bahar et al.

2007

Bahar et al.

2007

Bahar et al.

2007

Bahar et al. Elvik et al.

2007 2009

Elvik et al.

2009

International

CRF for going from unlit to lit CRF for going from unlit to lit CRF for going from unlit to lit

Elvik et al.

2009

International

Elvik et al.

2009

International

Elvik et al.

2009

International

CRF for going from unlit to lit

Elvik et al.

2009

International

Intersections CRF for going from unlit to lit CRF for going from unlit to lit CRF for going from unlit to lit Intersection CRF for going from unlit to lit CRF for going from unlit to lit

Elvik et al.

2009

International

Elvik et al.

2009

International

Elvik et al.

2009

International

Elvik et al.

2009

International

Austroads 2012 107

Effectiveness of Road Safety Engineering Treatments

Study Elvik et al.

Year 2009

Country International

Environment/Treatment CRF for going from unlit to lit

Reduction Night-time crashes on motorways Severity unknown Rear-end crashes CRF 20% (95% -36 +0) Single vehicle crashes CRF increase of 44% (95% -2 +110) Intersection/junction crashes CRF 41% (95% -64; -5) Night-time Increase level by up to double injury CRF 8% (95% CI -20; +6) Increase level by 25 times injury CRF 13% (95% CI -17; -9) Increase level by 5 times fatal CRF 50% (95% CI -79; +15) injury CRF 32% (95% CI -39; +25) Night-time crashes Rural two-lane roads, rural multilane highways, freeways, expressways, urban arterials, suburban arterials Injury CMF 0.71 (standard error unknown)

Elvik et al.

2009

International

CRF for increasing lighting levels

AASHTO

2010

USA

Base condition = unlit

Given the large number of studies on this topic, the information is summarised in the table below.
Environment/Treatment All Intersections 55% 41% 50% 55% Mid-blocks 50% 30% 45% 36% All 50% -30% -30% 33% 20% +25% 50% -40% 29% 33% 4%

Motorways

50% 4% (not significant) 50% 50% 40% 35% 22% 21% 40% 60% 60% 60%

Rural

8% 49% 31% 29% 59%

Urban Bridge Railway crossing

New installation

38% 25% 38% 50%

Austroads 2012 108

Effectiveness of Road Safety Engineering Treatments

Environment/Treatment Upgrade installation

Intersections 41%

Mid-blocks 30%

All 33% 25% 33% 50%

The effectiveness of lighting depends on a number of factors, including the luminance. The following sections provide simple crash reductions for the installation or upgrading of street lighting, but do not reflect issues such as the degree of lighting. Therefore a Crash Modification Function that took into account such factors would be more desirable than a simple percentage reduction. C.21.1 New Lighting All Sites A 35% reduction in night-time crashes for installing new lighting where no lighting was present, based on eight studies (one report rejected as an outlier). Given most of the results are consistent, there is a medium level of confidence. C.21.2 New Lighting Intersections A 50% reduction in night-time crashes at intersections with no lighting was based on the mean of four studies. Given the consistent results there is a high level of confidence in this figure. C.21.3 New Lighting Mid-blocks A 40% reduction in night-time crashes at mid-blocks based on the mean of four studies. The results are fairly consistent, so there is a high level of confidence in this figure. C.21.4 New Lighting Rural No CRF was calculated as there was one study which had results that were not significant, one figure that suggested an 8% reduction and one a 49% reduction. C.21.5 New Lighting Rural Intersections The CRF is 30% for night-time casualty reduction. There is medium confidence in this reduction as it is based on three studies suggesting fairly similar crash reductions. C.21.6 New Lighting Urban A 30% reduction in night-time crashes at urban sites based on two studies. Given there are just two studies, there is a low level of confidence. C.21.7 New Lighting Urban Intersections There is also an estimated 30% reduction at urban intersections. This figure is based on only two studies and has a low level of confidence. C.21.8 New Lighting Motorway Freeway Interchanges A reduction of 50% in night-time casualties is proposed. This is based on the mean of the three values (the 4% reduction was removed as it was not significant). There is a high level of confidence in this as all three studies suggest a 50% reduction in crashes. C.21.9 Railway Crossings A 60% reduction in night-time crashes is recommended, based on three studies that show the same result. There is high confidence in this figure.
Austroads 2012 109

Effectiveness of Road Safety Engineering Treatments

C.21.10 Upgrade Existing Lighting The benefit from improving existing lighting is estimated to be 35% (night-time casualties). There is a medium level of confidence in this figure. No information is provided on the level of improvement required to provide this benefit, but it is assumed to be significant. As a guide, more comprehensive guidance is provided in Elvik et al., and that research indicates that with a doubling of the lighting level there is an expected 8% decrease in night-time casualties; with between a two and five times increase there is a 13% decrease, and with a five times or more increase, there is an expected 32% decrease. C.21.11 Effect on Different Crash Types and Severities Various studies provided information on the expected crash reduction for different crash types (e.g. rear-end crashes) and severities (e.g. fatal). However, in most cases there is only one study that provides this information. Therefore, crash reduction factors have not been provided for these. C.21.12 Time of Day The vast majority of research provides CRFs for changes in casualties at night. There is no robust information on the effect of street lighting during day and night, and so no information has been provided on this. However, as would be expected, some of the research provides indications that the effect for both night and day combined is less than the effect for night-time alone. References assessed
AASHTO 2010, Highway safety manual, 1st edn, American Association of State Highway and Transportation Officials Washington, DC, USA. Agent, KR, Stamatiadis, N & Jones, S 1996, Development of accident reduction factors, report KTC-96-13, Kentucky Transportation Centre, University of Kentucky, Lexington, KY, USA. Andreassen, DC 1976, Vehicle conspicuity at night, Australian Road Research Board conference, 8th, Perth, Australian Road Research Board, Vermont South, Vic, vol.8, no.5, pp.26-43. Andreassen, DC 1989, Strategies for safety problems, research report ARR 163, Australian Road Research Board, Vermont South, Vic. Austroads 2009, Guide to road safety: part 8: treatment of crash locations, AGRS08/09, Austroads, Sydney, NSW. Bahar, G, Masliah, M, Wolff, R & Park, P 2007, Desktop reference for crash reduction factors, report FHWASA-07-015, Federal Highway Administration, Washington DC, USA. Bruneau, J, Morin, D & Pouliot, M 2001, Safety benefits of motorway lighting, Transportation Research Record, no. 1758, Transportation Research Board, Washington, pp.1-5. Bureau of Transport Economics (BTE) 2001, The black spot program 1996-2002: An evaluation of the first three years, report no. 104, Bureau of Transport Economics (BTE), Canberra, Australia. Corben, B, Deery, H, Mullan, N & Dyte, D 1997, The general effectiveness of countermeasures for crashes into fixed roadside objects, report 111, Monash University Accident Research Centre, Clayton, Vic. Creasey, T & Agent, KR 1985, Development of accident reduction factors, research report UKTRP-85-6, University of Kentucky, Kentucky Transportation Research Program, Lexington, KY.

Austroads 2012 110

Effectiveness of Road Safety Engineering Treatments

CTRE 2006, Safety impacts of street lighting at isolated rural intersections, part II, project web page, Center for Land Transport and Education (CTRE), Iowa State University, Ames, IA, USA, viewed 18 July 2011 <http://www.ctre.iastate.edu/research/detail.cfm?projectID=579>. Elvik, R, Hye, A, Vaa, T & Srensen, M 2009, The handbook of road safety measures, 2nd edn, Emerald Publishing Group, Bingley, UK. Ermer, DJ, Fricker, JD & Sinha, KC 1991, Accident reduction factors for Indiana, JHRP-91-11, School of Civil Engineering, Purdue University, Lafayette, IN, USA. Gorell, RSJ & Tootill, W 2001, Monitoring local authority road safety schemes using MOLASSES, report 512, TRL, Crowthorne, UK. Hall, RR & Fisher, AJ 1977, Measures of visibility and visual performance in road lighting, research report ARR 74, Australian Road Research Board, Vermont South, Vic. Isebrands, H, Hallmark, S, Hans, Z McDonald, T, Preston, H & Storm, R 2004, Safety impacts of street lighting at isolated rural intersections, part 2, year 1, draft report, Center for Land Transport and Education (CTRE), Iowa State University, Ames, IA, USA. Land Transport Safety Authority 1995, Accident investigation monitoring analysis version 3.0, Land Transport Safety Authority, Wellington, NZ. Land Transport Safety Authority 1997, Road lighting improvements, Land Transport Safety Authority, Wellington, NZ, viewed 18 July 2011, < http://www.nzta.govt.nz/resources/road-lightingimprovements/index.html >. Osram-Gec 1985, Road lighting: a proven countermeasure to road accidents, Lighting in Australia, vol.5, no.6, pp.35-6. Preston, H & Schoenecker, T 1999, Safety impacts of street lighting at isolated rural intersections, report MN/RC 1999-17, Minnesota Department of Transportation, Minnesota, USA. Sanderson, JT & Fildes, B 1984, Run-off-the-road accidents in rural areas, report TS84/6, Traffic and Safety, Royal Automobile Club of Victoria (RACV), Melbourne, Vic. Shen, J Rodriguez, A Gan, A & Brady, P 2004, Development and application of crash reduction factors: A state-of-the-practice survey of State Departments of Transportation, Transportation Research Board annual meeting, 83rd, Washington DC, TRB, Washington, DC, USA. Teale, G 1984, The evaluation of the effectiveness of low cost traffic engineering projects, consultant report CR 22, Office of Road Safety, Canberra, ACT.

Austroads 2012 111

INFORMATION RETRIEVAL
Austroads, 2012, Effectiveness of Road Safety Engineering Treatments, Sydney, A4, pp. 115. AP-R422-12 Keywords: Crash Reduction Factors, CRF, Crash Modification Factors, CMF, guide posts, chevron alignment markers, pavement markings, profile line marking, profile edge lines, profile centrelines, signs, regulatory signs, advisory signs, delineation, traffic signals, signal visibility, channelisation at intersections, splitter and median islands, grade separation, right-turn lane provision, extend right-turn lane, left-turn lane provision, lane width in tunnels, overtaking lanes, superelevation, impact attenuators, pedestrian treatments, transverse rumble strips, street closure, street lighting. Abstract: Previous Austroads research has identified that there is a lack of reliable information regarding the effectiveness of different road safety engineering treatments. This project addresses some of these gaps, providing information on the crash reduction effectiveness of a range of road safety engineering treatments.

Anda mungkin juga menyukai