Safety Effectiveness of Adding a Signal Head Per Lane in Wisconsin

Final Report: January 2018

Prepared by:

Kevin M. Scopoline, Wisconsin Department of Transportation

Kelly Greuel, AECOM

ii

ACKNOWLEDGEMENTS

The inspiration for this effort came from Joanna Bush. Without her dedication, this project would

not have succeeded. A special thanks is also extended to Brian Porter, Wisconsin DOT, for his

guidance and Jarrett Gates, Wisconsin DOT, for his data collection effort.

Jerry Roche, Federal Highway Administration, Office of Safety, was instrumental in the

completion of this effort. His enthusiasm and desire to help is greatly appreciated.

Thank you to VHB, specifically Frank Gross, Scott Himes, and Tim Harmon for their expertise

and many helpful comments and suggestions.

iii

Table of Contents

ACKNOWLEDGEMENTS .......................................................................................................... ii

EXECUTIVE SUMMARY ........................................................................................................... v

INTRODUCTION ....................................................................................................................... 1

LITERATURE REVIEW ............................................................................................................. 1

OVERVIEW OF METHODOLOGY ............................................................................................. 2

Simple Before-and-After Study ............................................................................................... 2

Empirical Bayes Before-and-After Study ................................................................................. 2

DATA COLLECTION ................................................................................................................. 4

Identification of Intersection Locations .................................................................................... 4

Crash Data Collection Process and Evaluation ....................................................................... 4

Geometric and Traffic Evaluation ............................................................................................ 5

Identifying Target Crash Types ............................................................................................... 5

Final Intersection Selections ................................................................................................... 5

RESULTS .................................................................................................................................. 6

Simple Before-and-After Analysis ........................................................................................... 6

Manner of Collision Analysis ............................................................................................... 6

Individual Injury Levels Analysis .......................................................................................... 7

Empirical Bayes Before-and-After Analysis ............................................................................ 8

Right-Angle Crashes ........................................................................................................... 8

Rear-End Crashes .............................................................................................................. 9

FHWA EB Methodology Review .........................................................................................11

CONCLUSIONS .......................................................................................................................11

Future Research ....................................................................................................................11

Appendix A: Intersection Data ................................................................................................. i

Appendix B: Empirical Bayes Results ..................................................................................... i

iv

List of Figures

Figure 1: EB Before-and-After Methodology ............................................................................... 3

Figure 2: Potential Site Locations ............................................................................................... 4

Figure 3: Final Intersection Selections ....................................................................................... 5

List of Tables

Table 1. Before and After Crash Data: Manner of Collision ........................................................ 6

Table 2. Before and After Crash Trends: Manner of Collision ..................................................... 7

Table 3. Before and After Crash Data: Individual Injury Levels ................................................... 7

Table 4. Before and After Crash Trends: Individual Injury Levels ............................................... 8

Table 5. EB Before and After Analysis: Right-Angle Crashes ..................................................... 9

Table 6. EB Before and After Analysis: Rear-End Crashes .......................................................10

v

EXECUTIVE SUMMARY

Traffic signal visibility is an important aspect of every signalized intersection. A primary crash

concern at signalized intersections is the number of intentional and unintentional factors that

cause drivers to run red-lights. Right-angle crashes where one vehicle violates the red signal

indication can result in serious injury crashes. Rear-end crashes can occur with a much higher

frequency but are typically less severe than right-angle crashes. Improving traffic signal visibility

is one countermeasure that may reduce red-lighting running and rear-end crashes.

The Wisconsin Department of Transportation (WisDOT) started installing monotube arm

assemblies in 2009 to provide the ability to install a signal head over the center of each travel

lane. WisDOT requires (since January 2012) monotube arm assemblies to be used on any

approach with two or more through lanes for reconstruction projects with state owned traffic

signals. This newer design standard does increase signal visibility at intersections but there are

tradeoffs. These larger signal poles and arms are non-breakaway structures which increases the

risk for vehicles that leave the roadway during a crash. Monotube assemblies can also increase

the complexity of design, create construction challenges, and add cost to a project.

This study analyzed 25 signalized intersections in Wisconsin that were reconstructed in 2009 or

2010 to provide signal heads per lane using monotube arm assemblies. All of the intersections

did receive other geometric improvements as these monotube installations were part of larger

reconstruction projects. Four years of before and after crash data was gathered and evaluated.

Geometric and traffic volume data was also collected for the before and after period at each

location. A simple before-and-after crash analysis was completed to evaluate target crash types

and individual injury severity levels for each intersection. An Empirical-Bayes (EB) analysis was

completed using Safety Performance Functions (SPFs) from the Highway Safety Manual (HSM)

to determine the safety benefits for right-angle and rear-end crashes.

Adding a signal head per lane with monotube arm assemblies showed a 23% decrease in right-

angle crashes. This study provided an opportunity to get local and relevant results for signal head

per lane installation for the state of Wisconsin. Signalized intersections in Wisconsin have always

put an emphasis on signal visibility through the use of other requirements such as backplates,

near-right signals, an overhead signal indication, and intersection street lighting. Signal head per

lane is another countermeasure that can increase safety by reducing right-angle crashes at

signalized locations, which are often severe.

1

INTRODUCTION

Traffic signal visibility is an important aspect of every signalized intersection. A primary crash

concern at signalized intersections is the number of intentional and unintentional factors that

cause drivers to run red-lights. Right-angle crashes where one vehicle violates the red signal

indication can result in serious injury crashes. Rear-end crashes can occur with a much higher

frequency but are typically less severe than right-angle crashes. Improving traffic signal visibility

is one countermeasure that may reduce red-lighting running and rear-end crashes.

Signal visibility from an adequate distance upstream can be improved by placement and number

of signal heads, size of signal display, and line of sight. The Manual on Uniform Traffic Control

Devices (MUTCD) provides guidance and requirements for signal head placement. One of the

additions to the 2009 MUTCD is guidance stating roadways with posted or statutory speed limits

or 85th-percentile speeds 45 mph or higher should have overhead primary signal faces centered

over each through lane. The MUTCD also recommends that this layout of signal faces should be

considered for any minor urban or suburban arterial street with speeds of less than 45 mph.

The Wisconsin Department of Transportation (WisDOT) started installing monotube arm

assemblies in 2009 to provide the ability to install a signal head over the center of each travel

lane. WisDOT requires (since January 2012) monotube arm assemblies to be used on any

approach with two or more through lanes for reconstruction projects with state owned traffic

signals. This newer design standard does increase signal visibility at intersections, but can create

other concerns. These bigger poles and arms are non-breakaway structures, which can cause

additional safety concerns. Monotube assemblies can also increase the complexity of design,

create construction challenges, and add cost to a project.

Wisconsin has other requirements in place to increase traffic signal visibility. This includes

requiring backplates, near-right signals, an overhead signal indication, and intersection street

lighting at all signalized intersections. With these other visibility measures already being

implemented, the safety benefits of installing monotube arm assemblies were unknown.

The objective of this study was to determine the safety benefits of installing monotube arm

assemblies at signalized intersections in Wisconsin. Currently, WisDOT owns 980 traffic signals

throughout the state and approximately a third of these signalized intersections already have been

upgraded to include monotube arm assemblies with more planned in the future. The goal of this

study is to develop unbiased, comprehensive evaluation methodologies, and quantify the safety

of signal head per lane through the use of monotubes.

LITERATURE REVIEW

Few research studies have been completed regarding the effectiveness of signal head per lane.

Previous studies and crash modification factors (CMFs) from the Highway Safety Manual (HSM)

or CMF Clearinghouse mainly focus on increasing signal visibility by upgrading signal lens sizes,

installing backplates, adding reflective backplates, or increasing the number of signal heads.

These studies provided evidence that increased signal visibility can provide safety benefits, such

as reducing the crash frequency and crash severity. However, these studies do not represent a

2

typical Wisconsin signalized intersection and as a result, are not well suited to estimate the

benefits of installing a signal head per lane in Wisconsin.

Signal head use per lane has become a standard practice, especially on higher posted speed

roadways. Many Wisconsin intersections have been upgraded with monotubes to accomplish

increased visibility by placing signal heads centered over each lane. It is unclear how this has

impacted the safety of these intersections. This indicates the need for further research in this

area using data from Wisconsin.

OVERVIEW OF METHODOLOGY

Simple Before-and-After Study

A simple before-and-after study is a basic comparison of crash data before and after a specific

treatment is installed. It calculates the reduction in crashes due to the implementation of a

treatment. A positive value for the change in safety or a ratio less than one indicates that the

treatment provided a reduction in crashes at the site.

∆ 𝑆𝑎𝑓𝑒𝑡𝑦 = 𝐵𝑤𝑖𝑡ℎ𝑜𝑢𝑡 − 𝐴𝑤𝑖𝑡ℎ

𝐶𝑀𝐹 =𝐴𝑤𝑖𝑡ℎ

𝐵𝑤𝑖𝑡ℎ𝑜𝑢𝑡

Where:

A = the number of crashes in the after period

B = the number of crashes in the before period

Although simple before-and-after studies are beneficial in determining changes in crashes, this

type of comparison assumes that the change in crashes is solely attributed to the safety

improvement and that everything else remains constant before and after installing the treatment.

This creates potential issues and biases that are common to these types of studies. Before-and-

after studies do not account for bias due to trends associated with traffic volumes, driver behavior

and regression-to-the-mean (RTM) effects.

RTM bias is a statistical phenomenon where locations with abnormally high or low crashes will

regress towards the mean frequency of crashes during the following years. An extreme

observation in crashes will usually be followed by a less extreme observation in crashes without

any intervention. If bias due to RTM effects are not addressed, a study may overestimate or

underestimate the safety effect of a treatment.

Empirical Bayes Before-and-After Study

The Highway Safety Manual provides several methods of analysis for completing a safety

effectiveness evaluation. One method is the Empirical Bayes (EB) before-and-after study. The

EB method is used to compare crash frequencies at a group of sites before and after a treatment

3

is implemented. EB uses a safety performance function (SPF) to predict the average crash

frequency. By using SPFs, the EB method addresses the bias due to changes in traffic volumes,

driver behavior, and RTM effects. An Empirical Bayes before-and-after study estimates the

expected number of crashes that would have occurred at a location had there been no treatment

and compares that with the number of observed crashes after the treatment. This allows for a

more reliable estimate of the expected number of crashes without the treatment than the simple

before-and-after analysis.

This study followed the HSM, Chapter 9 methodology to complete the EB evaluation and identify

appropriate SPFs. Table 12-10 of Section 12.6.2 in the HSM provides applicable SPF coefficients

for multiple-vehicle collisions at intersections that could be used to predict crashes for this study.

The final analysis was performed using the urban arterial, four-legged signalized intersection

SPFs. Those SPFs were calibrated using the data collected in this study and the procedure

described in the HSM.

As outlined in Chapter 9 of the HSM, the EB methodology is shown in Figure 1 and has the

following main steps:

1. Estimation of the expected crash frequency in the before period

2. Estimation of the expected crash frequency in the after period

3. Estimation of treatment effectiveness

4. Estimate of precision of the treatment effectiveness

Figure 1: EB Before-and-After Methodology

Observed

Crashes

Expected

Crashes

Expected Crashes

w/o Treatment

Observed

Crashes

# C

rash

es

Before Period without

Treatment

After Period with

Treatment

Time

SPF

Predicted

Crashes

Predicted

Crashes

CRF

4

DATA COLLECTION

Identification of Intersection Locations

The study began with determining possible intersection

locations that could be used for the analysis. To complete a

meaningful before-and-after comparison, a minimum of

three years of before and after period crash data is required.

Therefore, intersections with upgrades to signal head per

lane with the use of monotubes were only considered if the

intersection was reconstructed in the years 2009 and 2010.

Based on input from the WisDOT traffic signal staff, 42

possible intersection locations were selected for initial

review. Figure 2 shows the locations of the 42 signalized

intersections.

All intersections upgraded with this safety improvement

were part of larger reconstruction projects and received

other geometric and signal improvements. In addition, all of

the possible sites were in either the WisDOT Northeast (NE)

Region or the Southeast (SE) Region. Most of the first

monotube installations happened in the SE Region because

traffic volumes are higher and there are more multilane facilities. Although this does not provide

a good representation of the whole state, it does provide an analysis using intersection locations

that are similar in roadway type, geometry, and traffic volumes.

Crash Data Collection Process and Evaluation

Crash data for the signalized intersection locations was collected from WisTransPortal for the four

years before and after the installation of signal head per lane with the use of monotubes. Crashes

that occurred during the construction year of the intersection were excluded.

Due to the importance of gathering accurate crash data, independent crash analyses were

completed by both WisDOT and AECOM. Information was collected for any crash that was coded

to the particular intersection of interest. An initial review of the crash data was completed to filter

out crashes that were not associated with the intersection. This included eliminating crashes that

were animal related or non-intersection related crashes that were more than 500 feet away from

the intersection. Wisconsin crash report forms (MV4000) were obtained for the filtered crash data

and a detailed manual review of each crash was completed by both parties. This included

reviewing the police narratives and diagrams to determine which crashes were at and related to

the signalized intersection while also verifying the manner of collision for each crash.

Working meetings between WisDOT and AECOM were conducted after both parties completed

the independent crash analysis. The crash data was finalized for the safety analysis for before

and after the installation of monotubes. Before and after crash diagrams were also created for all

intersections.

Figure 2: Potential Site Locations

5

Geometric and Traffic Evaluation

All traffic signal plans were reviewed and data was collected before and after the installation of

signal head per lane with the use of monotubes. This information provided details to changes

that were made at the intersection in addition to signal head per lane. Important design features

collected included posted speed limits, lane configurations, left-turn phasing, presence of flashing

yellow arrows (FYAs), right-turn control, and pedestrian/bicycle accommodations.

Historical annual average daily traffic (AADT) information was collected from WisDOT

Roadrunner (interactive traffic count map) and from WisDOT Southeast Region staff. No turning

movement traffic counts were conducted as a separate task for this study. AADT for the mainline

ranged from 12,000 to 35,000 and side road AADT ranged from 1,000 to 25,000.

Identifying Target Crash Types

All the intersections were improved as part of a larger reconstruction project. Several other

improvements were made to each of the intersections in addition to the monotube installation. To

isolate the safety benefits of the signal head per lane upgrade, the specific crash types most

impacted by signal visibility were identified as target crash types. Right-angle crashes where

drivers disregarded traffic control and rear-end crashes were the target crash types for this study.

Angle crashes that involved left-turning traffic that may have misjudged a gap of an opposing

through vehicle, and right-turn rear-end crashes, were not included as a target crash type for this

analysis.

Final Intersection Selections

After all the data was collected at each of the intersections,

final intersection selection was determined. 25 of the original

42 intersections were selected for the safety evaluation. The

goal was to choose intersections with similar attributes such

as roadway type, speeds, and traffic volumes. Intersections

were eliminated for various reasons:

• Unique geometry

• Interchanges

• Rural intersections

• AADT’s missing or outside of study range

• Number of intersection approaches

• Change in traffic control

The final 25 signalized intersections selected included sites from two Wisconsin counties:

Waukesha and Milwaukee County. This provided similar roadway types, volumes, speeds, and

driver behavior since all study sites were taken from the same area of the state. Figure 3 shows

the location of the 25 final signalized intersection selections.

Figure 3: Final Intersection Selections

6

RESULTS

Simple Before-and-After Analysis

Two simple before-and-after evaluations were completed for this study. One analysis evaluated

manner of collision for all crash types, right-angle crashes, and rear-end crashes and the other

evaluated individual injury levels. These analyses did not take into consideration any RTM

effects, but were useful in evaluating crash types and crash severity levels.

Manner of Collision Analysis

Table 1 shows the observed crash statistics for all 25 study signalized intersections in the before

and after period broken out by the targeted crash types.

Table 1. Before and After Crash Data: Manner of Collision

Signalized Intersection

Before Period (crashes in 4 years)

After Period (crashes in 4 years)

Right-Angle

Rear-End

ALL Right-Angle

Rear-End

ALL

STH 241 & Layton Ave 2 33 64 7 25 52

STH 24 & 60th St 3 7 18 2 4 15

STH 36 & Grange Ave 4 3 26 3 5 20

STH 241 & College Ave 5 24 54 4 24 46

STH 24 & Grange Ave 2 9 16 2 7 12

STH 36 & Ramsey Ave 2 10 28 1 2 12

STH 24 & Morgan Ave 2 5 12 1 5 10

STH 241 & Grange Ave 13 7 33 6 14 28

STH 38 & Puetz Rd 4 11 61 5 27 51

STH 100 & 91st St 12 20 51 4 10 25

STH 241 & Edgerton Ave 6 5 29 9 10 31

STH 241 & Cold Spring Rd 8 14 36 5 10 36

STH 24 & Howard Ave 0 5 8 1 3 4

STH 24 & 68th St 5 6 18 2 4 13

STH 100 & Range Lane Rd 1 11 26 3 2 10

STH 100 & 107th St 7 7 41 9 6 38

STH 241 & Ramsey Ave 4 9 23 2 7 20

STH 24 & 84th St 7 0 20 3 2 11

STH 36 & College Ave 3 5 24 2 3 13

STH 36 & Drexel Ave 5 7 20 1 4 23

STH 100 & Kildeer Ct 5 5 19 4 5 18

STH 241 & Parnell Ave 5 8 19 1 8 15

STH 83 & Heritage Dr 7 25 43 4 27 43

STH 83 & Golf Rd 6 7 43 5 15 50

STH 83 & Hillside Dr 4 15 56 1 8 43

TOTAL 122 258 788 87 237 639

7

Table 2 shows the number of locations with increase, no change, or decrease in crashes between

the before and after periods and the overall crash reduction over all intersections broken out by

manner of collision. For all targeted crash types, the number of locations with reduced crashes

was greater than the number of locations with increased crashes. Right-angle crashes saw the

largest overall reduction in crashes, with a 29% reduction.

Table 2. Before and After Crash Trends: Manner of Collision

Crash Type Number of Locations

Overall Crash Reduction Over All Intersections

Increase No Change Decrease # Reduced % Reduction

Right-Angle 6 1 18 35 29%

Rear-End 7 4 14 21 8%

All Crashes 3 2 20 149 19%

Individual Injury Levels Analysis

Table 3 shows the observed crash statistics for all 25 study signalized intersections in the before

and after period broken out by individual injury levels. Individual injury levels are broken out by

K, A, B, C, and property damage only (PDO). The crash severity listed is for the worst level of

severity to life and property as a result of a particular crash.

Table 3. Before and After Crash Data: Individual Injury Levels

Signalized Intersection

Before Period (crashes in 4 years)

After Period (crashes in 4 years)

K A B C PDO Total K A B C PDO Total

STH 241 & Layton Ave 0 1 7 20 36 64 0 1 3 12 36 52

STH 24 & 60th St 0 0 2 4 12 18 0 0 1 1 13 15

STH 36 & Grange Ave 0 1 4 7 14 26 0 0 4 4 12 20

STH 241 & College Ave 0 3 5 16 30 54 0 1 8 12 25 46

STH 24 & Grange Ave 0 0 0 3 13 16 0 1 1 3 7 12

STH 36 & Ramsey Ave 0 3 2 13 10 28 0 0 3 2 7 12

STH 24 & Morgan Ave 0 1 1 2 8 12 0 0 4 1 5 10

STH 241 & Grange Ave 0 2 1 11 19 33 1 1 2 10 14 28

STH 38 & Puetz Rd 0 0 8 11 42 61 0 0 5 6 40 51

STH 100 & 91st St 0 1 3 20 27 51 0 0 4 9 12 25

STH 241 & Edgerton Ave 0 1 4 10 14 29 0 1 5 12 13 31

STH 241 & Cold Spring Rd 0 1 4 15 16 36 1 0 5 16 14 36

STH 24 & Howard Ave 0 0 1 3 4 8 0 0 1 0 3 4

STH 24 & 68th St 0 0 3 7 8 18 0 0 2 5 6 13

STH 100 & Range Lane Rd 0 3 1 5 17 26 0 1 2 2 5 10

STH 100 & 107th St 0 0 5 5 31 42 0 0 7 9 22 38

STH 241 & Ramsey Ave 0 0 6 5 12 23 0 1 4 6 9 20

8

STH 24 & 84th St 0 0 2 6 12 20 0 2 1 0 8 11

STH 36 & College Ave 0 0 3 7 14 24 0 1 1 3 8 13

STH 36 & Drexel Ave 0 2 5 7 6 20 0 0 4 7 12 23

STH 100 & Kildeer Ct 0 0 1 4 14 19 0 0 2 4 12 18

STH 241 & Parnell Ave 0 0 0 5 14 19 0 1 1 3 10 15

STH 83 & Heritage Dr 0 0 2 12 29 43 0 1 2 7 33 43

STH 83 & Golf Rd 0 1 2 4 36 43 0 0 3 8 39 50

STH 83 & Hillside Dr 0 0 4 14 38 56 0 1 4 6 32 43

TOTAL 0 20 76 216 476 788 2 13 79 148 397 639

Table 4 shows the number of locations with increase, no change, or decrease in crashes between

the before and after periods and the overall crash reduction over all intersections broken out by

individual injury levels. There were no fatal (K) crashes in the before period and two in the after

period. The two fatal crashes in the after period were not a result of hitting a monotube structure.

For injury level A and C crashes the number of locations with reduced crashes was greater than

the number of locations with increased crashes. Injury level B crashes saw a greater number of

locations with increased crashes in the after period compared to the before period. For PDO

crashes, the number of locations with decreases in crashes was greater than the number of

locations with increases in crashes.

Table 4. Before and After Crash Trends: Individual Injury Levels

Crash Severity

Number of Locations Overall Crash Reduction Over

All Intersections

Increase No Change Decrease # Reduced % Reduction

K 2 23 0 -2 ---

A 7 8 10 7 35%

B 13 4 8 -3 -4%

C 5 3 17 68 31%

PDO 4 1 20 79 17%

Empirical Bayes Before-and-After Analysis

An EB before-and-after analysis was completed for the targeted crash types: right-angle and rear-

end. This analysis addresses the bias due to RTM effects and traffic volumes and utilized SPFs

found in the HSM.

Right-Angle Crashes

Table 5 shows the results for the EB analysis for right-angle crashes. For each intersection, the

EB expected crashes in the after period without treatment is shown, along with the actual crashes

in the before and after period, and the safety effectiveness of the treatment at each site in the

form of an odds ratio and as a percentage of the crash change at each site.

9

Table 5. EB Before and After Analysis: Right-Angle Crashes

Signalized Intersection

Before Period After Period

Odds2

Ratio

Safety3

Effectiveness (%)

Observed Total Crashes

Expected Crashes w/o Treatment

Observed Total

Crashes

STH 241 & Layton Ave 2 2.4 7 2.92 -192.2

STH 24 & 60th St 3 3.4 2 0.58 41.6

STH 36 & Grange Ave 4 2.8 3 1.08 -8.4

STH 241 & College Ave 5 6.6 4 0.61 39.0

STH 24 & Grange Ave 2 2.2 2 0.89 10.8

STH 36 & Ramsey Ave 2 2.2 1 0.46 54.2

STH 24 & Morgan Ave 2 2.5 1 0.41 59.4

STH 241 & Grange Ave 13 10.1 6 0.59 40.5

STH 38 & Puetz Rd 4 3.2 5 1.58 -57.5

STH 100 & 91st St 12 9.3 4 0.43 57.0

STH 241 & Edgerton Ave 6 5.8 9 1.55 -55.4

STH 241 & Cold Spring Rd 8 8.1 5 0.62 38.2

STH 24 & Howard Ave 0 1.2 1 0.85 14.6

STH 24 & 68th St 5 5.0 2 0.40 60.2

STH 100 & Range Lane Rd 1 1.9 3 1.57 -56.6

STH 100 & 107th St 7 6.4 9 1.40 -40.3

STH 241 & Ramsey Ave 4 5.2 2 0.39 61.5

STH 24 & 84th St 7 4.0 3 0.75 25.1

STH 36 & College Ave 3 3.1 2 0.65 35.2

STH 36 & Drexel Ave 5 4.0 1 0.25 75.2

STH 100 & Kildeer Ct 5 4.4 4 0.91 8.8

STH 241 & Parnell Ave 5 5.8 1 0.17 82.9

STH 83 & Heritage Dr 7 5.4 4 0.74 26.2

STH 83 & Golf Rd 6 3.6 5 1.39 -39.2

STH 83 & Hillside Dr 4 4.2 1 0.24 76.1

TOTAL 122 112.7 87 0.77 22.8

The results of the right-angle EB before-and-after analysis show that 18 locations (72%) showed

a decrease in crash frequency and 7 locations (28%) showed an increase. Overall, there was a

23% decrease in right-angle crashes across all locations, significant at the 95% confidence level

(standard error of 10.0%).

Rear-End Crashes

Table 6 shows the results for the EB analysis for rear-end crashes. For each intersection, the EB

expected crashes in the after period without treatment is shown, along with the actual crashes in

Notes:

1) SPF intersection type used for all intersections was U-A-4SG-M

2) Odds Ratio (ORi) = Nobserved,A / Nexpected,A

3) Safety Effectivenessi = 100 x (1 – ORi)

10

the before and after period, and the safety effectiveness of the treatment at each site in the form

of an odds ratio and as a percentage of the crash change at each site.

Table 6. EB Before and After Analysis: Rear-End Crashes

Signalized Intersection

Before Period After Period

Odds2

Ratio

Safety3

Effectiveness (%)

Observed Total Crashes

Expected Crashes w/o Treatment

Observed Total

Crashes

STH 241 & Layton Ave 33 21.1 25 1.18 -18.2

STH 24 & 60th St 7 7.5 4 0.53 46.9

STH 36 & Grange Ave 3 3.2 5 2.52 -152.3

STH 241 & College Ave 24 28.7 24 0.84 16.3

STH 24 & Grange Ave 9 7.3 7 0.96 4.1

STH 36 & Ramsey Ave 10 8.1 2 0.25 75.3

STH 24 & Morgan Ave 5 5.5 5 0.91 9.3

STH 241 & Grange Ave 7 7.3 14 1.92 -92.5

STH 38 & Puetz Rd 11 8.0 27 3.38 -238.1

STH 100 & 91st St 20 16.9 10 0.59 40.8

STH 241 & Edgerton Ave 5 5.9 10 1.69 -69.2

STH 241 & Cold Spring Rd 14 15.7 10 0.64 36.3

STH 24 & Howard Ave 5 4.6 3 0.65 35.4

STH 24 & 68th St 6 7.6 4 0.52 47.5

STH 100 & Range Lane Rd 11 8.7 2 0.23 77.0

STH 100 & 107th St 7 8.3 6 0.72 27.8

STH 241 & Ramsey Ave 9 10.7 7 0.65 34.7

STH 24 & 84th St 0 1.5 2 1.37 -37.3

STH 36 & College Ave 5 5.6 3 0.54 46.2

STH 36 & Drexel Ave 7 6.9 4 0.58 41.9

STH 100 & Kildeer Ct 5 5.1 5 0.98 2.3

STH 241 & Parnell Ave 8 9.8 8 0.81 18.8

STH 83 & Heritage Dr 25 23.0 27 1.17 -17.2

STH 83 & Golf Rd 7 5.0 15 3.01 -201.4

STH 83 & Hillside Dr 15 14.8 8 0.54 45.8

TOTAL 258 246.8 237 0.97 2.8

The results of the rear-end EB before-and-after analysis show that 17 locations (68%) showed a

decrease in crash frequency and 8 locations (32%) showed an increase. Overall, there was a 3%

decrease in rear-end crashes across all locations, which was not significant at 90% confidence

level (standard error of 8.3%). One potential reason why the reduction for rear-end crashes is

not significant may be due to an under reporting of crashes.

Notes:

1) SPF intersection type used for all intersections was U-A-4SG-M

2) Odds Ratio (ORi) = Nobserved,A / Nexpected,A

3) Safety Effectivenessi = 100 x (1 – ORi)

11

FHWA EB Methodology Review

This study was supported by the Federal Highway Association (FHWA) through the Every Day

Counts (EDC) program. As part of this support, the EB methodology was reviewed for correctness

of application as well as providing guidance on some of the finer points of the process since this

was the first study conducted by WisDOT staff. Some helpful advice included: using a proportion

of total multi-vehicle crashes to get the number of right angle and rear-end crashes; to only use

the last year of before data when calibrating in order to account for RTM effects, and categorizing

data to make sure the analysis can be performed.

CONCLUSIONS

This study analyzed 25 Wisconsin signalized intersections that were reconstructed in 2009 or

2010 to provide signal heads per lane through the use of monotube arm assemblies. All the

intersections received other geometric improvements as these monotube installations were part

of larger reconstruction projects. Four years of before and after crash data was gathered and

evaluated. Geometric and traffic volume data was also collected for the before and after period.

A simple before-and-after crash analysis was completed to evaluate target crash types and

individual injury severity levels for each intersection. An EB analysis was completed using SPFs

from the HSM to determine the safety benefits for right-angle and rear-end crashes.

Overall, the analysis shows that signal head per lane with the use of monotube arm assemblies

reduces right-angle and rear-end crashes. Rear-end crashes were reduced in the simple before-

after analysis but results from the EB analysis were not statistically significant at a 90% confidence

level. Right-angle crashes saw the largest improvement with an average crash reduction of 23%.

Installing a signal head per lane provides better visibility for drivers and provides a clearer

message on multi-lane approaches. Even though these larger monotube structures are non-

breakaway and are more complex to design and construct, the study showed that there is a safety

benefit to installing a signal head per lane at signalized intersections.

This study provided an opportunity to get local results for signal head per lane installation for the

state of Wisconsin. Signalized intersections in Wisconsin have always put an emphasis on signal

visibility through the use of other requirements such as backplates, near-right signals, an

overhead signal indication, and intersection street lighting. Signal head per lane is another

countermeasure that can increase safety and reduce right-angle, higher severity crashes in

Wisconsin and other states around the country.

Future Research

This research has shown that adding a signal head per lane with monotube assemblies has led

to improvements in traffic safety in Wisconsin. This initial study evaluated the first monotube

structure installations in the state. Most of these installations were in the southeast part of the

state and were part of larger reconstruction projects. Since these initial implementations, more

monotube arm assemblies have been installed throughout all WisDOT regions. In addition,

implementation of monotube arm assemblies has been completed as standalone traffic signal

improvement projects.

12

Future research should consider intersections in other WisDOT regions where adding a signal

head per lane was the only improvement made at the intersection. Evaluating rear end crashes

with additional data may help to gain insight into the effects this treatment has on those crashes.

A - i

Appendix A: Intersection Data

Table A-1: List of Study Intersections

Intersection Year Built Municipality County Included

in EB? If not included in EB, why?

STH 50 & CTH P 2010 Twin Lakes Kenosha No Changed control type

STH 241 & Layton 2009 Milwaukee Milwaukee Yes

STH 24 & 60th 2009 Milwaukee Milwaukee Yes

STH 36 & Grange 2010 Greendale Milwaukee Yes

STH 241 & College Ave 2010 Milwaukee Milwaukee Yes

STH 24 & Grange 2009 Hales Corner Milwaukee Yes

STH 36 & Ramsey 2010 Greendale Milwaukee Yes

STH 24 & Morgan 2009 Greenfield Milwaukee Yes

STH 241 & Grange 2010 Greendale Milwaukee Yes

STH 24 & Cold Springs 2009 Greenfield Milwaukee No Monotube not on every approach

IH 43 & Good Hope Rd - East 2010 Glendale Milwaukee No AADT outside SPF range

IH 43 & Good Hope Rd - West 2010 Glendale Milwaukee No AADT outside SPF range

I-43 EB & STH 241 (South) 2009 Greenfield Milwaukee No Monotube not on every approach

STH 38 & Puetz 2010 Oak Creek Milwaukee Yes

STH 100 & 91st St 2010 Milwaukee Milwaukee Yes

STH 241 & Edgerton 2010 Milwaukee Milwaukee Yes

STH 241 & Coldspring 2010 Greenfield Milwaukee Yes

STH 24 & Howard 2009 Milwaukee Milwaukee Yes

STH 24 & 68th 2009 Greenfield Milwaukee Yes

STH 100 & Range Line 2009 River Hills Milwaukee Yes

STH 100 & 107th St 2010 Milwaukee Milwaukee Yes

STH 241 & Ramsey 2010 Milwaukee Milwaukee Yes

STH 24 & 84th 2009 Greenfield Milwaukee Yes

STH 36 & College 2010 Greendale Milwaukee Yes

STH 36 & Drexel 2010 Franklin Milwaukee Yes

STH 100 & Kildeer 2009 Brown Deer Milwaukee Yes

STH 36 & 76th St 2010 Greendale Milwaukee No Monotube not on every approach

IH 94 SB & College (East) 2009 Milwaukee Milwaukee No AADT outside SPF range

IH 94 SB & College (West) 2009 Milwaukee Milwaukee No AADT outside SPF range

STH 241 & Parnell 2010 Milwaukee Milwaukee Yes

STH 47 & Evergreen Dr 2010 Grand Chute Outagamie No Missing AADT information

STH 11 & STH 32 2010 Mount Pleasant Racine No Monotube not on every approach

STH 20 & STH 36_83 2010 Rochester Racine No Changed control type

STH 36_83 & CTH W 2010 Burlington Racine No Changed control type

STH 36_83 & CTH D 2010 Rochester Racine No Changed control type

STH 83 & IH 94 WB 2010 Delafield Waukesha No AADT outside SPF range

STH 83 & Heritage 2010 Delafield Waukesha Yes

STH 83 & IH 94 EB 2010 Delafield Waukesha No AADT outside SPF range

STH 83 & Golf Rd 2010 Delafield Waukesha Yes

STH 83 & Hillside 2010 Delafield Waukesha Yes

A - ii

Table A-2: Intersection Crash Data by Severity

Intersection Before Period Crashes by Severity After Period Crashes by Severity

K A B C PDO Total K A B C PDO Total

STH 241 & Layton 0 1 7 20 36 64 0 1 3 12 36 52

STH 24 & 60th 0 0 2 4 12 18 0 0 1 1 13 15

STH 36 & Grange 0 1 4 7 14 26 0 0 4 4 12 20

STH 241 & College Ave 0 3 5 16 30 54 0 1 8 12 25 46

STH 24 & Grange 0 0 0 3 13 16 0 1 1 3 7 12

STH 36 & Ramsey 0 3 2 13 10 28 0 0 3 2 7 12

STH 24 & Morgan 0 1 1 2 8 12 0 0 4 1 5 10

STH 241 & Grange 0 2 1 11 19 33 1 1 2 10 14 28

STH 38 & Puetz 0 0 8 11 42 61 0 0 5 6 40 51

STH 100 & 91st St 0 1 3 20 27 51 0 0 4 9 12 25

STH 241 & Edgerton 0 1 4 10 14 29 0 1 5 12 13 31

STH 241 & Coldspring 0 1 4 15 16 36 1 0 5 16 14 36

STH 24 & Howard 0 0 1 3 4 8 0 0 1 0 3 4

STH 24 & 68th 0 0 3 5 10 18 0 0 2 5 6 13

STH 100 & Range Line 0 3 1 5 17 26 0 1 2 2 5 10

STH 100 & 107th St 0 0 5 5 31 41 0 0 7 9 22 38

STH 241 & Ramsey 0 0 7 6 11 24 0 1 4 6 9 20

STH 24 & 84th 0 0 2 6 13 21 0 2 1 0 8 11

STH 36 & College 0 0 3 7 14 24 0 1 1 3 8 13

STH 36 & Drexel 0 2 5 7 6 20 0 0 4 7 12 23

STH 100 & Kildeer 0 0 1 4 14 19 0 0 2 4 12 18

STH 241 & Parnell 0 0 0 5 14 19 0 1 1 3 10 15

STH 83 & Heritage 0 0 2 12 28 42 0 1 2 7 33 43

STH 83 & Golf Rd 0 1 2 4 37 44 0 0 3 8 39 50

STH 83 & Hillside 0 0 4 14 38 56 0 1 4 6 32 43

A - iii

Table A-3: Intersection Crash Data by Type

Intersection

Before Period Crashes by Type After Period Crashes by Type

Single Vehicle

Right Angle

Left Turn

Rear End

Right Turn

Other Multi - Vehicle

Total Single

Vehicle Right Angle

Left Turn

Rear End

Right Turn

Other Multi - Vehicle

Total

STH 241 & Layton 8 2 8 33 3 10 64 2 7 2 25 5 11 52

STH 24 & 60th 2 3 3 7 0 3 18 2 2 3 4 2 2 15

STH 36 & Grange 2 4 14 3 0 3 26 3 3 3 8 2 2 21

STH 241 & College Ave 4 5 13 24 2 6 54 7 4 2 24 3 6 46

STH 24 & Grange 0 2 2 9 3 0 16 0 2 3 7 0 0 12

STH 36 & Ramsey 4 2 10 10 1 1 28 1 1 4 2 3 1 12

STH 24 & Morgan 2 2 1 5 0 2 12 2 1 2 5 0 0 10

STH 241 & Grange 2 13 9 7 0 2 33 0 6 5 14 0 3 28

STH 38 & Puetz 3 4 35 11 3 5 61 7 5 5 27 5 2 51

STH 100 & 91st St 6 12 8 20 1 4 51 2 4 3 10 0 6 25

STH 241 & Edgerton 2 6 15 5 0 1 29 3 9 6 10 0 3 31

STH 241 & Coldspring 1 8 9 14 1 3 36 3 5 11 10 0 7 36

STH 24 & Howard 2 0 1 5 0 0 8 0 1 0 3 0 0 4

STH 24 & 68th 3 5 2 6 0 2 18 4 2 2 4 0 1 13

STH 100 & Range Line 1 1 6 11 0 7 26 1 3 2 2 1 1 10

STH 100 & 107th St 3 7 14 7 3 7 41 3 9 14 6 4 2 38

STH 241 & Ramsey 3 4 2 9 1 4 24 6 2 3 7 0 3 21

STH 24 & 84th 2 7 7 0 4 0 21 2 3 3 2 0 1 11

STH 36 & College 5 3 9 5 0 2 24 1 2 5 3 0 2 13

STH 36 & Drexel 0 5 6 7 0 2 20 4 1 7 4 1 6 23

STH 100 & Kildeer 3 5 4 5 1 1 19 1 4 5 5 1 2 18

STH 241 & Parnell 2 5 4 8 0 0 19 0 1 4 8 1 1 15

STH 83 & Heritage 1 7 0 25 6 3 42 2 4 0 27 4 6 43

STH 83 & Golf Rd 5 6 7 7 14 5 44 5 5 7 15 14 4 50

STH 83 & Hillside 7 4 14 15 6 10 56 12 1 10 8 10 2 43

A - iv

Table A-4: Intersection AADTs

Intersection

Before Period AADT After Period AADT

Year 1 Year 2 Year 3 Year 4 Year 1 Year 2 Year 3 Year 4

Major Minor Major Minor Major Minor Major Minor Major Minor Major Minor Major Minor Major Minor

STH 241 & Layton 36,200 22,100 35,958 22,383 35,717 22,667 35,475 22,950 34,992 23,517 34,750 23,800 32,140 24,136 29,537 24,478

STH 24 & 60th 14,141 12,612 14,058 12,869 13,978 13,132 13,900 13,400 14,250 14,833 14,425 15,550 14,833 14,517 15,242 13,483

STH 36 & Grange 16,183 13,062 16,513 13,328 16,850 13,600 16,483 14,031 15,750 14,900 15,617 13,817 15,483 12,733 15,350 11,650

STH 241 & College Ave

25,825 15,017 25,950 14,742 26,075 14,467 28,117 14,192 32,200 13,642 32,812 13,583 33,431 13,525 34,055 13,467

STH 24 & Grange 14,636 7,482 14,934 7,635 15,239 7,791 15,550 7,950 15,917 7,117 16,100 6,700 16,283 6,867 16,467 7,033

STH 36 & Ramsey 17,505 6,460 17,862 6,492 18,227 6,525 18,347 6,433 18,600 6,250 17,717 6,253 16,833 6,257 15,950 6,262

STH 24 & Morgan 13,185 5,238 13,454 5,344 13,729 5,453 14,009 5,565 14,900 5,854 15,350 6,000 15,983 5,900 16,617 5,800

STH 241 & Grange 27,252 4,525 27,382 4,618 27,518 4,712 28,740 4,750 31,716 5,751 29,771 5,969 27,981 5,753 26,197 6,189

STH 38 & Puetz 26,172 9,813 26,706 9,750 27,251 9,688 27,807 9,625 24,713 9,500 22,904 9,200 22,379 9,558 21,853 9,679

STH 100 & 91st St 28,407 9,114 29,200 9,300 28,242 9,300 27,283 9,300 27,750 9,867 29,175 10,433 30,600 11,000 32,025 11,267

STH 241 & Edgerton 30,196 4,642 30,820 4,571 31,450 4,500 33,167 4,467 36,600 4,400 31,600 4,271 26,600 4,143 21,600 4,016

STH 241 & Coldspring

30,009 3,483 29,319 3,667 28,628 3,850 28,804 4,183 29,156 4,850 29,332 5,183 30,003 5,517 30,680 5,850

STH 24 & Howard 16,450 5,553 15,650 5,666 14,850 5,782 14,050 5,900 14,320 6,867 14,460 7,350 14,602 6,967 14,747 6,583

STH 24 & 68th 15,391 5,300 14,941 5,083 14,494 4,867 14,050 4,650 18,650 4,783 20,950 4,850 20,921 4,983 20,895 5,117

STH 100 & Range Line

24,010 3,392 24,500 3,383 25,000 3,375 23,667 3,367 21,000 3,350 21,167 3,192 21,333 3,034 21,500 2,877

STH 100 & 107th St 20,965 11,708 21,606 11,947 22,109 11,429 22,615 10,913 24,317 11,517 25,508 12,633 26,700 13,750 26,792 9,617

STH 241 & Ramsey 29,690 3,799 29,479 3,748 29,269 2,698 29,058 3,753 30,434 3,857 31,153 3,910 31,776 3,989 32,412 4,068

STH 24 & 84th 13,930 5,600 14,214 5,975 14,504 6,350 14,800 6,725 14,767 7,475 14,750 7,850 13,917 7,308 13,083 6,767

STH 36 & College 15,687 3,709 15,892 3,785 16,100 3,868 16,667 3,850 17,800 3,657 17,300 3,677 16,800 3,769 16,300 3,861

STH 36 & Drexel 12,819 6,195 13,081 6,321 13,348 6,450 13,620 6,817 15,825 7,550 17,687 8,000 18,119 8,450 18,550 8,900

STH 100 & Kildeer 26,988 5,948 27,539 6,070 28,101 6,194 27,317 6,320 25,117 5,927 25,367 5,730 25,060 5,544 24,754 5,357

STH 241 & Parnell 27,132 1,859 27,142 1,897 27,158 1,936 28,260 1,975 31,216 1,929 29,260 1,906 27,460 1,944 25,665 1,983

STH 83 & Heritage 18,071 3,200 18,440 2,933 18,816 2,667 19,200 2,400 21,800 2,497 23,100 2,547 22,633 2,598 22,167 2,650

STH 83 & Golf Rd 31,248 5,232 31,885 5,339 32,536 5,448 33,200 5,559 25,667 5,668 21,900 5,750 21,533 5,832 21,767 5,915

STH 83 & Hillside 17,506 5,150 17,863 4,583 18,228 4,017 18,600 3,450 18,533 3,589 18,500 3,661 17,883 3,734 17,267 3,809

B - i

Appendix B: Empirical Bayes Results

Table B-1: Empirical Bayes Results for Right Angle Crashes

Intersection Prop. Factor

Predicted Before Crashes Adj.

Nexp

Before

Predicted After Crashes Adj.

Nexp

After Odds Ratio

Safety Effect.

Var() Y1 Y2 Y3 Y4

Total

Y1 Y2 Y3 Y4 Total

STH 241 & Layton Ave 0.17 2.3 2.2 2.2 2.2 9.0 0.2 3.5 1.6 1.6 1.5 1.4 6.0 0.7 2.4 2.9 -192.2 1.3

STH 24 & 60th St 0.17 0.9 0.9 0.9 0.9 3.6 0.4 3.3 0.9 0.9 1.0 1.0 3.8 1.1 3.4 0.6 41.6 2.1

STH 36 & Grange Ave 0.17 0.6 0.7 0.7 0.7 2.6 0.5 3.3 0.6 0.6 0.5 0.5 2.2 0.8 2.8 1.1 -8.4 1.2

STH 241 & College Ave 0.17 1.1 1.1 1.1 1.1 4.3 0.4 4.7 1.4 1.5 1.5 1.5 5.9 1.4 6.6 0.6 39.0 5.7

STH 24 & Grange Ave 0.17 0.9 0.9 0.9 0.9 3.5 0.4 2.6 0.7 0.7 0.8 0.8 3.0 0.8 2.2 0.9 10.8 1.1

STH 36 & Ramsey Ave 0.17 0.6 0.7 0.7 0.7 2.7 0.5 2.3 0.7 0.6 0.6 0.6 2.5 0.9 2.2 0.5 54.2 1.0

STH 24 & Morgan Ave 0.17 0.6 0.6 0.7 0.7 2.6 0.5 2.3 0.7 0.7 0.7 0.7 2.8 1.1 2.5 0.4 59.4 1.3

STH 241 & Grange Ave 0.17 1.6 1.6 1.6 1.7 6.6 0.3 11.2 1.6 1.5 1.4 1.3 5.9 0.9 10.1 0.6 40.5 6.5

STH 38 & Puetz Rd 0.17 1.2 1.2 1.3 1.3 5.0 0.3 4.3 1.0 0.9 0.9 0.9 3.7 0.7 3.2 1.6 -57.5 1.5

STH 100 & 91st St 0.17 1.6 1.7 1.6 1.6 6.6 0.3 10.5 1.3 1.4 1.5 1.6 5.8 0.9 9.3 0.4 57.0 5.9

STH 241 & Edgerton Ave 0.17 1.8 1.8 1.9 2.0 7.5 0.3 6.4 2.2 1.9 1.5 1.2 6.8 0.9 5.8 1.6 -55.4 3.9

STH 241 & Cold Spring Rd 0.17 1.1 1.1 1.1 1.1 4.5 0.4 6.7 1.3 1.3 1.4 1.4 5.4 1.2 8.1 0.6 38.2 6.2

STH 24 & Howard Ave 0.17 0.7 0.7 0.7 0.6 2.7 0.5 1.3 0.6 0.6 0.6 0.6 2.4 0.9 1.2 0.9 14.6 0.5

STH 24 & 68th St 0.17 0.7 0.7 0.7 0.7 2.8 0.5 3.9 0.8 0.9 0.9 0.9 3.5 1.3 5.0 0.4 60.2 3.3

STH 100 & Range Line Rd 0.17 1.4 1.5 1.5 1.4 5.8 0.3 2.5 1.1 1.1 1.1 1.1 4.5 0.8 1.9 1.6 -56.6 1.0

STH 100 & 107th St 0.17 0.8 0.8 0.9 0.9 3.4 0.4 5.4 0.9 1.0 1.1 1.0 4.0 1.2 6.4 1.4 -40.3 4.3

STH 241 & Ramsey Ave 0.17 1.7 1.7 1.6 1.7 6.6 0.3 4.7 1.7 1.8 1.8 1.9 7.3 1.1 5.2 0.4 61.5 4.1

STH 24 & 84th St 0.17 0.6 0.6 0.7 0.7 2.6 0.5 4.8 0.6 0.6 0.5 0.5 2.2 0.8 4.0 0.7 25.1 1.7

STH 36 & College Ave 0.17 0.7 0.7 0.7 0.7 2.9 0.5 2.9 0.8 0.8 0.7 0.7 3.0 1.0 3.1 0.6 35.2 1.7

STH 36 & Drexel Ave 0.17 0.6 0.6 0.7 0.7 2.6 0.5 3.8 0.6 0.7 0.7 0.7 2.8 1.1 4.0 0.2 75.2 2.1

STH 100 & Kildeer Ct 0.17 1.7 1.7 1.8 1.7 6.9 0.3 5.5 1.4 1.4 1.4 1.3 5.5 0.8 4.4 0.9 8.8 2.5

STH 241 & Parnell Ave 0.17 1.3 1.3 1.3 1.4 5.4 0.3 5.1 1.7 1.6 1.5 1.4 6.1 1.1 5.8 0.2 82.9 4.5

STH 83 & Heritage Dr 0.17 0.4 0.4 0.4 0.4 1.6 0.6 3.7 0.6 0.6 0.6 0.6 2.4 1.5 5.4 0.7 26.2 3.1

STH 83 & Golf Rd 0.17 1.1 1.1 1.1 1.1 4.4 0.4 5.4 0.8 0.7 0.7 0.7 2.9 0.7 3.6 1.4 -39.2 1.5

STH 83 & Hillside Dr 0.17 0.8 0.8 0.7 0.7 3.0 0.5 3.5 0.9 0.9 0.9 0.9 3.5 1.2 4.2 0.2 76.1 2.7

Odds Ratio: 0.772 Standard Error: 0.101 Unbiased Odds Ratio: 0.768 Standard Error of Safety Effectiveness: 10.1

Overall Unbiased Safety Effectiveness: 23.2 Statistical Comparison: 2.31 Variance of Unbiased Odds Ratio: 0.01 Significant at the 95% confidence level

B - ii

Table B-2: Empirical Bayes Results for Rear End Crashes

Intersection Prop. Factor

Predicted Before Crashes Adj.

Nexp

Before

Predicted After Crashes Adj.

Nexp

After Odds Ratio

Safety Effect.

Var() Y1 Y2 Y3 Y4

Total

Y1 Y2 Y3 Y4 Total

STH 241 & Layton Ave 0.36 4.8 4.7 4.7 4.7 18.9 0.1 31.3 3.4 3.4 3.1 2.9 12.8 0.7 21.1 1.2 -18.2 12.6

STH 24 & 60th St 0.36 1.9 1.9 1.9 1.9 7.6 0.3 7.2 2.0 2.0 2.0 2.1 8.0 1.1 7.5 0.5 46.9 5.9

STH 36 & Grange Ave 0.36 1.3 1.4 1.4 1.4 5.5 0.3 3.8 1.2 1.2 1.1 1.1 4.6 0.8 3.2 2.5 -152.3 1.8

STH 241 & College Ave 0.36 2.2 2.2 2.2 2.4 9.1 0.2 20.7 3.1 3.1 3.2 3.2 12.6 1.4 28.7 0.8 16.3 30.9

STH 24 & Grange Ave 0.36 1.8 1.8 1.9 1.9 7.5 0.3 8.6 1.6 1.6 1.6 1.6 6.3 0.8 7.3 1.0 4.1 4.6

STH 36 & Ramsey Ave 0.36 1.4 1.4 1.4 1.4 5.6 0.3 8.6 1.4 1.4 1.3 1.2 5.3 0.9 8.1 0.2 75.3 5.2

STH 24 & Morgan Ave 0.36 1.3 1.4 1.4 1.4 5.6 0.3 5.2 1.4 1.4 1.5 1.6 5.9 1.1 5.5 0.9 9.3 4.0

STH 241 & Grange Ave 0.36 3.4 3.4 3.5 3.6 13.9 0.2 8.1 3.4 3.2 3.0 2.8 12.5 0.9 7.3 1.9 -92.5 5.5

STH 38 & Puetz Rd 0.36 2.6 2.6 2.7 2.7 10.6 0.2 10.9 2.1 1.9 1.9 1.8 7.7 0.7 8.0 3.4 -238.1 4.7

STH 100 & 91st St 0.36 3.5 3.6 3.5 3.3 13.9 0.2 19.0 2.8 3.0 3.2 3.4 12.3 0.9 16.9 0.6 40.8 12.7

STH 241 & Edgerton Ave 0.36 3.8 3.9 4.0 4.2 15.8 0.1 6.5 4.6 3.9 3.2 2.6 14.4 0.9 5.9 1.7 -69.2 4.6

STH 241 & Cold Spring Rd 0.36 2.4 2.4 2.3 2.4 9.4 0.2 13.0 2.7 2.8 2.9 3.0 11.4 1.2 15.7 0.6 36.3 14.9

STH 24 & Howard Ave 0.36 1.5 1.5 1.4 1.3 5.7 0.3 5.2 1.3 1.3 1.3 1.3 5.1 0.9 4.6 0.6 35.4 2.8

STH 24 & 68th St 0.36 1.6 1.5 1.4 1.4 5.9 0.3 6.0 1.7 1.9 1.9 1.9 7.5 1.3 7.6 0.5 47.5 6.8

STH 100 & Range Line Rd 0.36 3.0 3.1 3.2 3.0 12.3 0.2 11.2 2.4 2.4 2.4 2.4 9.5 0.8 8.7 0.2 77.0 5.6

STH 100 & 107th St 0.36 1.7 1.8 1.8 1.8 7.1 0.3 7.0 2.0 2.1 2.3 2.1 8.4 1.2 8.3 0.7 27.8 7.2

STH 241 & Ramsey Ave 0.36 3.6 3.6 3.3 3.5 14.0 0.2 9.8 3.7 3.8 3.9 4.0 15.3 1.1 10.7 0.7 34.7 9.9

STH 24 & 84th St 0.36 1.3 1.3 1.4 1.4 5.5 0.3 1.7 1.2 1.2 1.1 1.0 4.6 0.8 1.5 1.4 -37.3 0.8

STH 36 & College Ave 0.36 1.5 1.5 1.5 1.6 6.1 0.3 5.3 1.7 1.6 1.6 1.5 6.4 1.0 5.6 0.5 46.2 4.1

STH 36 & Drexel Ave 0.36 1.3 1.4 1.4 1.5 5.6 0.3 6.5 1.3 1.5 1.5 1.6 5.8 1.1 6.9 0.6 41.9 5.0

STH 100 & Kildeer Ct 0.36 3.6 3.7 3.8 3.7 14.7 0.1 6.4 3.0 3.0 2.9 2.8 11.7 0.8 5.1 1.0 2.3 3.5

STH 241 & Parnell Ave 0.36 2.8 2.8 2.8 2.9 11.3 0.2 8.6 3.6 3.3 3.1 2.9 13.0 1.1 9.8 0.8 18.8 9.2

STH 83 & Heritage Dr 0.36 0.9 0.9 0.9 0.9 3.5 0.4 15.8 1.2 1.3 1.3 1.3 5.0 1.5 23.0 1.2 -17.2 19.2

STH 83 & Golf Rd 0.36 2.2 2.3 2.4 2.4 9.3 0.2 7.5 1.7 1.5 1.5 1.5 6.2 0.7 5.0 3.0 -201.4 2.6

STH 83 & Hillside Dr 0.36 1.6 1.6 1.6 1.6 6.3 0.3 12.5 1.9 1.9 1.9 1.8 7.5 1.2 14.8 0.5 45.8 12.4

Odds Ratio: 0.972 Standard Error: 0.083 Unbiased Odds Ratio: 0.969 Standard Error of Safety Effectiveness: 8.3

Overall Unbiased Safety Effectiveness: 3.1 Statistical Comparison: 0.37 Variance of Unbiased Odds Ratio: 0.01 No significant at the 90% confidence level