Md Mozahidul Islam_MS Thesis

i

A COMPREHENSIVE ASSESSMENT OF POSSIBLE LINKS BETWEEN DIGITAL

ADVERTISING BILLBOARDS AND TRAFFIC SAFETY

by

MD MOZAHIDUL ISLAM

VIRGINIA P. SISIOPIKU, COMMITTEE CHAIR

IAN E. HOSCH

ANDREW SULLIVAN

A THESIS

Submitted to the graduate faculty of The University of Alabama at Birmingham,

in partial fulfillment of the requirements for the degree of

Master of Science

BIRMINGHAM, ALABAMA

2015

ii

Copyright by

MD MOZAHIDUL ISLAM

2015

iii

A COMPREHENSIVE ASSESSMENT OF POSSIBLE LINKS BETWEEN DIGITAL

ADVERTISING BILLBOARDS AND TRAFFIC SAFETY

MD MOZAHIDUL ISLAM

MASTERS OF SCIENCE IN CIVIL ENGINEERING

ABSTRACT

Advertising billboards are a common roadside object and a very efficient medium

of outside advertising. For years, static billboards have been adopted by the billboard

advertising companies. Most recently, a sizeable portion of the regular (static) billboards

have been digitized to convey more information to the drivers, thus raising questions

about their potential impact on traffic safety. Frequently changing images on digital

advertising billboards may compel more glances, and sequential messages may hold

drivers‟ gazes longer until the entire message is read.

Earlier studies sponsored by billboard advertising companies did not report

statistically significant correlations between the crash occurrences and the presence of

digital billboards. Some other studies tried to show potential relationship between

diminished attention caused by digital billboards to crashes but suffered from

methodological problems and did not succeed due to lack of sufficiently reliable manner.

So, there is an ongoing debate surrounding this issue and a need for an objective and

reliable evaluation to determine if the presence of digital billboards really distracts

driver‟s attention or not and, if distraction occurs then to what extent.

To bridge these gaps, this thesis studied the correlation between the presence of

digital billboards and traffic safety through a. literature review, b. driver questionnaire

survey, and c. crash data analysis. The literature review involved a comprehensive

review and synthesis of findings from existing studies on digital advertising billboards,

iv

driver distraction and traffic safety. The survey of road users focused on the

development of a questionnaire survey that was used to survey Alabama drivers and

document road user‟s perceptions and attitudes related to roadside advertising billboards.

The crash analysis involved an analysis of historical crashes along selected interstate

routes in Alabama to determine if the presence of digital billboards has an impact on

crash occurrence. The goal was to compare the frequency of crashes within the billboard

area of influence to crash frequencies at adjacent comparison sites. The crash rate by type

and severity has also been determined at u/s (or, study site or influence zone) and d/s (or,

control site or non-influence zone) of digital billboard locations.

The findings of this thesis are expected to assist policy makers to better

understand the effect of digital billboards from the safety viewpoint. If required,

amendments for digital billboard size and location guidelines may be introduced for

improving the overall safety of road users.

Keywords: Digital advertising billboards, traffic safety, driver distraction, crash analysis

v

DEDICATION

This thesis is dedicated to my parents who have given me moral and mental

support.

vi

ACKNOWLEDGMENTS

All praise to Almighty Allah, the most Gracious and most Merciful.

The author would like to express his sincere appreciation and gratitude to his

supervisor, Dr. Virginia Sisiopiku, Associate Professor, Department of Civil,

Construction, and Environmental Engineering, University of Alabama at Birmingham

(UAB), for her continuous guidance, invaluable suggestions and affectionate

encouragement at all stage of this study. Without her valuable direction and cordial

assistance, this research work could never be materialized. The author‟s debt to her is

immense.

The author is thankful to Dr. Ozge Cavusoglu, PhD, University of Alabama at

Birmingham for her cordial help in crash data analysis.

The author is grateful to the authors of different articles mentioned in the

reference which proved to be very helpful throughout the whole thesis work.

The author is indebted to Mr. Mostafa Emeira who helped with identifying and

taking images of digital billboard locations.

Finally, the author is willing to show his solemn gratitude to his parents for their

continuous support and motivation throughout the thesis work.

vii

TABLE OF CONTENTS

Page

ABSTRACT ....................................................................................................................... iii

DEDICATION .....................................................................................................................v

ACKNOWLEDGMENTS ................................................................................................. vi

LIST OF TABLES ............................................................................................................. ix

LIST OF FIGURES .............................................................................................................x

LIST OF ABBREVIATIONS ........................................................................................... xii

CHAPTER

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

1.1 Background ...................................................................................................................1

1.2 Objectives ...................................................................................................................2

1.3 Scope ...................................................................................................................3

1.4 Organization ...................................................................................................................4

2 LITERATURE REVIEW .................................................................................................5

2.1 General ...........................................................................................................................5

2.2 Literature Synthesis Studies ...................................................................................5

2.3 Crash Studies .........................................................................................................9

2.4 Summary ..............................................................................................................11

3 SURVEY OF DRIVER‟S PERCEPTIONS....................................................................12

3.1 General .................................................................................................................12

3.2 Methods................................................................................................................12

3.3 Analysis................................................................................................................13

3.4 Results ..................................................................................................................13

3.5 Summary ..............................................................................................................23

viii

4 CRASH RECORDS ANALYSIS: DATA COLLECTION CONSIDERATIONS ........24

4.1 General .................................................................................................................24

4.2 Approach ..............................................................................................................24

4.3 Summary ..............................................................................................................30

5 CRASH DATA ANALYSIS: METHODOLOGY AND RESULTS .............................31

5.1 General .................................................................................................................31

5.2 Data Analysis Procedure ......................................................................................31

5.3 Results ..................................................................................................................33

5.3.1 Analysis of Crash Records Trends ..........................................................33

5.3.2 Crash Analysis Results ............................................................................36

5.4 Discussion ............................................................................................................40

5.5 Summary ..............................................................................................................40

6 CONCLUSIONS AND RECOMMENDATIONS .........................................................41

6.1 Summary of Research ..........................................................................................41

6.2 Implications for Practice ......................................................................................41

6.3 Limitations and Future Research .........................................................................42

LIST OF REFERENCES ...................................................................................................44

APPENDICES

A CHI-SQUARE TEST RESULT FROM ONLINE QUESTIONNAIRE

SURVEY ........................................................................................................46

B AGGREGATE CRASH ANALYSIS .............................................................50

ix

LIST OF TABLES

Table Page

3.1 Aggregate Response from Online Questionnaire Survey ......................................14

3.2 Chi-Square Test Result for Age Groups ................................................................22

3.3 Chi-Square Test Result for Male and Female ........................................................23

4.1 List of Alabama Study (u/s) and Control (d/s) Sections ........................................29

5.1 Crash Summary Statistics at the Digital Billboard Locations

(Aggregate Value) ..................................................................................................37

5.2 Summary Statistics by Crash Type ........................................................................38

5.3 Summary Statistics by Crash Injury Severity ........................................................39

x

LIST OF FIGURES

Figure Page

3.1 Number of Respondents with Age Class ...............................................................15

3.2 Perception on Distraction by Billboards with Respect to Age ..............................16

3.3 Perception on More Distraction Potential of Digital Billboard with

Respect to Age .......................................................................................................16

3.4 More Likeliness to Read Digital Billboard with Respect to Age ..........................17

3.5 Long Glance at Digital Billboard with Respect to Age .........................................18

3.6 Slow Down to Digital Billboard with Respect to Age ...........................................19

3.7 Use of Information from Digital Billboard with Respect to Age ..........................20

3.8 Perception on Restriction on Location of Digital Billboards with

Respect to Age .......................................................................................................20

3.9 Perception on Restriction on Size and Number of Digital Billboards with

Respect to Age .......................................................................................................21

4.1 Steps Associated with the Alabama Crash Rate Study ..........................................25

4.2 Typical Study Location ..........................................................................................26

4.3 Spatial Representation of Study Locations (On County-by-County Basis)...........28

4.4 Location ID 7 on I-459 in Bessemer (Jefferson County) .......................................29

4.5 Location ID 8 on I-20/59 in Bessemer (Jefferson County) ....................................30

5.1 Aggregate Crash Frequency by Year .....................................................................33

xi

5.2 Crash Frequency at DBB Influence Zones and Control Segments by Year ..........34

5.3 Crash Frequency at Study Sites by Month .............................................................35

5.4 Crash Frequency at Study Sites by Day of the Week ............................................35

5.5 Crash Frequency at Study Sites by Time of the Day .............................................36

xii

LIST OF ABBREVIATIONS

AADT Average Annual Daily Traffic

ALDOT Alabama Department of Transportation

CARE Critical Analysis Reporting Environment

CMS Changeable Message Sign

CR Crash Rate

DBB Digital Billboard

d/s Downstream

EBM Empirical Bayes Method

FHWA Federal Highway Administration

NCHRP National Cooperative Highway Research Program

OAAA Outdoor Advertising Association of America

u/s Upstream

VDOT Virginia Department of Transportation

VMT Vehicle Miles of Travel

1

CHAPTER 1

INTRODUCTION

1.1 Background

Roadside advertising billboards are used for advertisement of various products

and services and are meant to attract drivers‟ attention to the message or information

conveyed by the billboards. According to the Outdoor Advertising Association of

America (OAAA), there were over 365,000 unique billboard faces in the United States in

2013 (Outdoor Advertising Association of America [OAAA], 2013).

Roadside advertising billboards can be either static or digital. Static billboards

show the same message for an extended period of time (typically days). They are the

traditional type of outdoor advertising and the most commonly used type of advertising

billboards in the United States. The digital billboards (DBBs) were introduced in the

recent years and utilize light-emitting diode (LED) technology to display multiple

messages one after another that are updated using computer input. Because DBBs flash

images every four to ten seconds (Copeland, 2010), a single board can advertise to far

more clients than a traditional board, making them an attractive advertisement option.

Thus, despite the fact that DBBs are initially more expensive to build compared to their

static counterparts, over time they prove to be cost-effective. While static billboards are

still dominant, digital billboards are a fast growing sector of the outdoor advertising

market (OAAA, 2013).

The increased number and sophistication of DBBs raises questions about their

potential impact on traffic safety. As an advertising medium, DBBs purposely

encouraging drivers to shift their attention away from the driving task. Moreover, DBBs

2

brightness may be especially problematic at night and may affect the driver‟s ability to

observe changes in the surrounding environment such as brake lights or signal changes.

Also, frequently changing images may compel more glances, and sequential messages

may hold drivers‟ gazes longer until the entire message is read. Lastly, targeted messages

that promote interactivity with the driver are particularly troublesome as they are

hypothesized to be distracting to the driver (Sisiopiku et al., 2013).

Earlier studies sponsored by billboard advertising companies stated that the

presence of digital billboards does not cause a change in driver behavior in terms of

visual behavior, speed maintenance, or lane keeping (Lee et al., 2007). In the past,

attempts have been made to show that driver's diminished attention could result in more

crashes in the vicinity of such billboards, but because of the methodological problems of

these studies this has never been done in a sufficiently reliable manner (Institute for Road

Safety Research, Roadside Advertising and Information, 2013). Due to the growing

debate on this issue, an objective evaluation is needed to determine if the presence of

digital billboards really distracts driver‟s attention or not and, if distraction occurs then to

what extent.

1.2 Objectives

The overall objective of this research is to investigate potential relationships

between presence of digital billboards and traffic safety (driver distraction). In order to

meet this objective the following steps are needed:

Consider the distribution of digital billboards along the interstate routes in

Alabama and select study sections

3

Analyze spatial representation of the billboards and crashes at digital billboard

locations (both study and control sections)

Conduct spatial analysis (distribution of crashes based on distance from digital

billboards)

Seek for relationship between crash occurrence and presence of digital billboards.

1.3 Scope

The scope of this study is to investigate potential relationship between digital

billboards and traffic safety. This will be done by using statistical analysis of crash rates

at billboard locations. Detailed data will consist of digital billboard data and historical

crash data over a period of five years. The data will provide the necessary information to

identify the location of digital billboards along limited access facilities (interstates) in

Alabama and also the number, type and severity of crashes in those locations. In the first

stage, the digital billboard data will be used to identify digital billboard locations and

make a spatial representation of them using ArcGIS 10.1. This will enable the selection

of candidate study locations. Candidate locations will be further evaluated on the basis of

selection criteria. Locations that meet such criteria will be retained for further analysis.

The study location identification and selection process will be described in Chapter 2.

At the second stage sufficient historical crash data will be obtained from available

sources. These data will be processed to get exact crash counts at selected study locations

and control locations. After that the crash rates will be determined and comparisons

between crash rates at DBB areas of influence and corresponding control locations will

be performed. This will allow the observation on any changes in crash numbers and

pattern of crashes (e.g., type, and severity) over space due to the presence of digital

4

billboards. Traffic data (i.e., Annual Average Daily Traffic, AADT) will be also taken

into consideration to make the analysis more logical since the increase in traffic may

cause more crashes naturally.

1.4 Organization

This thesis comprises of seven chapters to illustrate the methodology for

achieving the aforementioned objectives and associated results and conclusions. The

thesis is organized as below.

Chapter 1 provides an introduction and discusses the context of the study

Chapter 2 focuses on the synthesis of the literature related to this topic by

comparison of different approaches, cross referencing, recommendation and

necessary citations

Chapter 3 describes questionnaire survey data collection and analysis result prior

to the collection of digital billboard and crash data

Chapter 4 presents digital billboard and crash data collection procedure and

description of the study sites

Chapter 5 provides crash data processing, analysis and results with respect to

aggregate crash rate, crash type and crash injury severity

Chapter 6 discusses the findings of the thesis and gives directions for future

research

5

CHAPTER 2

LITERATURE REVIEW

2.1 General

Digital advertising billboards are a commonplace feature along interstates,

highways and roadways. This chapter provides a detailed literature review on the

presence of digital billboards and its potential impact on driver distraction or traffic

safety. Next, some past researches about spatial and temporal analysis of crash data will

be discussed. This will summarize earlier findings and provide insights for the analysis

will be done in the subsequent chapters.

2.2 Literature Synthesis Studies

Several literature review and meta-analyses exist on the subject of outdoor

advertising and driver distraction. A few of such studies were funded by non-neutral

sources, so the results reported should be considered with discernment.

Wallace (2003) used meta-analysis to investigate whether or not there is a serious

safety risk caused by features in the external driving environment. After twelve selected

studies were analyzed, Wallace concluded that there seemed to be an association between

crash rates and billboards at intersections. The only one of the twelve studies that showed

no relationship between crashes and advertising billboard signs was performed on a

stretch of road that contained no intersections. Secondly, the author reported a possible

correlation between crash rates, billboard signs, and sharp bends after long stretches of

road. Thirdly, concerning the first two conclusions, the evidence was largely situation-

specific. Wallace also stated that many studies have shown that billboards had little to no

6

impact on driver safety, but still many indicated outdoor advertising can be a serious

threat to road safety. Wallace concluded that the subject is under-researched and thus new

research is needed to combine past knowledge with current practices paving the way for

additional studies in the recent years (Wallace, 2003).

In a parallel effort, Coetzee (2003) reviewed and summarized the findings from

six previous crash studies (Minnesota Department of Highways Field Study, 1951; Iowa

State College Field Study, 1951; Michigan State Highway Department Field Study 1952;

Faustman, California Route Field Study, 1961). Among the studies considered was a

1951 study done by the Minnesota Department of Highways that is known as one of the

first advertising billboard-driver safety studies. It reported that in a sample of 713

crashes, intersections with 4 or more billboards had a crash rate 3 times higher than at

intersections with no billboards. The same year, Iowa State University evaluated crash

rates immediately upstream and immediately downstream of billboards and found that

crash rates upstream were double the rates downstream. In 1952, the Michigan State

Highway Department found that billboards had no effect on crash rates, although it was

concluded that illuminated signs exhibited a correlation with crash locations. Crash rates

reported in another study found that the addition of one billboard at a given location

resulted in a 12.3% increase in crashes, while the addition of 5 billboards resulted in a

61.7% increase in crashes (Coetzee, 2003).

A report facilitated by FHWA reviewed potential concerns on driving safety

associated with digital billboards. Research on driver performance, state regulatory

practices, tri-vision signs, literature review, roadway characteristics‟ relationship to driver

distraction, driver characteristics‟ relationship to driver safety, and the legibility of

7

Changeable Message Signs (CMSs) were included in the report. Also included was a

section describing research needs on the subject (Farbry et al., 2001).

A similar report released by the FHWA in 2009 described how the recent

emergence of DBBs along U.S. roadways has caused a need for a reevaluation of current

legislation and regulation for controlling outdoor advertising. Driver distraction emerged

as a chief concern. This report consisted of earlier published work, research of applicable

research methods and techniques, and recommendations for future research (Molino et

al., 2009).

In 2009, Wachtel issued a report under National Cooperative Highway Research

Program (NCHRP) Project 20-7 (256) to help state and local governments establish

guidelines for outdoor advertising signs. Included in the report is a) an identification of

human factors related to digital outdoor advertising, b) an investigation into existing

regulations on outdoor advertising in both the U.S and abroad, and c) a review of the

current literature on the subject. The studies reviewed in the report were separated into

two distinct categories: i.e., neutral research and industry-funded studies. Because the

technology of DBBs is relatively novel, more research on the subject has transpired in

recent years; out of the 150 studies cited in the report, 20 occurred in the last decade.

Wachtel highlighted several successful regulations to serve as models for other entities to

consider. He also concluded that the relationship between DBBs and driver distraction is

very complex. The dynamic nature of field studies in roadway corridors presents many

challenges to achieve objective research, and laboratory studies have a limited

relationship with reality. One suggestion to remedy this problem would be to design a

study that combines the validity of a field study with the control of a laboratory setting.

8

Moreover, the fact that DBBs are quickly adapting and evolving as technology advances

makes offering guidelines on the issue even more challenging. Adding to the complexity

is the fact that industry-funded studies may include biased conclusions. However, despite

the convolution of the issue, Wachtel concludes that that there is enough of a solid and

growing body confirming that roadside advertising attracts drivers‟ eyes away from the

road for discernibly unsafe periods of time. It remains to be seen whether or not the

combination of existing, in progress, and future research is sufficient for the alteration of

current industry standards (Wachtel, 2009).

The U.S. Sign Council issued a response to the 2009 Wachtel report that is critical

of Wachtel‟s work, claiming that his recommendations were limited in scope, and

unnecessarily criticized studies that use scientific methods. The Council, which is funded

by the advertising industry, also reported that only a small percentage of the literature

reviewed in the report involved field studies, and that the author invited the reader to

“take a circuitous path around existing studies” on digital billboards and driver distraction

in order to reach a conclusion that billboards are a distraction (Crawford, 2010).

In a follow-up report, Wachtel focused on how digital billboards distract U.S.

drivers. The report suggested that DBBs cause drivers to be less observant of stopping

cars ahead of them, and contribute to vehicle drifting into adjacent lanes. The report also

offered suggestions on ways to control the effects of digital advertising, which include

controlling the lighting of the signs, keeping the signs simple, and prohibiting message

sequencing (Wachtel, 2011).

9

2.3 Crash Studies

Most of the previous crash studies involve spatial and/or temporal analysis of

crash data. Spatial analysis looks at the variation of crash rates with distance from the

digital billboard (DBB). The temporal analysis incorporates the variation of crash rates

from time to time. Sometimes it is useful to compare crash rates before and after

conversion of billboards (from static to digital).

In a 2010 report, Tantala and Tantala examined the statistical relationship

between digital billboards and traffic safety in Albuquerque, New Mexico. Analysis of

traffic and crash data was conducted for a 7-year period on local roads near 17 DBBs.

Each billboard contained one digital plane that was converted from traditional signage

between 2006 and 2007. First, the researchers reviewed the frequency of crashes near the

billboards before and after conversion to digital. Ranges analyzed in the study included

0.2, 0.4, 0.6, 0.8, and 1.0 miles both upstream and downstream of each sign. Also, time of

day and age of driver dynamics were factored into the study. Secondly, the researchers

performed a spatial analysis to investigate the potential correlation between the locations

of billboards and crashes. The results of the study indicated that the 17 digital billboards

in Albuquerque have no significant relationship with auto crashes. Specifically, crash

rates near the digital boards showed a 0.3% decrease in crash rate within 0.6 miles of the

signs over a period of six years. Furthermore, the spatial component of the study found

no significant clustering of crashes in the vicinity of billboard sites (Tantala and Tantala,

2010a).

Tantala and Tantala (2010) also examined the statistical correlation between

digital billboards and crash data in Henrico County and Richmond, Virginia. The study

10

analyzed crash data in the vicinity of 14 digital billboards. Data sources included

municipal police departments, Henrico County, and the Virginia Department of

Transportation (VDOT). The structure of the research was similar to the Albuquerque

study; 7 years of crash data (approximately 40,000 crashes) were examined at sites near

the selected billboards, which were converted from conventional to digital faces during

the time period of 2006 to 2009. Once again, temporal and spatial components were

investigated within ranges of a half mile upstream and downstream of the billboards. An

Empirical Bayes Method (EBM) analysis was utilized to approximate the number of

crashes that could be expected without the presence of signs. Results indicated that digital

billboards in the Richmond area had no statistically significant relationship with crash

occurrence. The evaluation of the EBM analysis indicated that the actual number of

crashes in each location was consistent with what would be expected with or without the

institution of digital billboards (Tantala and Tantala, 2010b).

In 2012, Yannis and colleagues conducted a statistical analysis applied on road

sites in the Athens, Greece metropolitan area. The goal of the research was to investigate

the relationship between the placement and removal of advertising signs and the related

occurrence of road incidents. Crash data from the test sites were obtained from the

Hellenic Statistical Authority database and analyzed. The analysis showed no correlation

between road crashes and advertising signs in any of the nine sites examined (Yannis et

al., 2012).

In another research effort, the city of Toronto requested an investigation of the

effects of billboards and safety on three downtown intersections and one expressway.

Five distinct studies were carried out: a. an eye movement study; b. a conflict study at

11

intersection approaches; c. a speed study; d. crash analysis, and e. a public questionnaire

survey. Results from the first study indicated that drivers glanced at video signs 50% of

the time, with 20% of all glances lasting more than 0.75 seconds. The conflict study

revealed that significantly more braking occurred near intersections in the presence of

video signs. The third study confirmed that driving speed decreased and speed variance

increased after the billboard sign was installed. In the fourth study, there was no

substantial increase in crashes near signed approaches. Lastly, 65% of those surveyed

believed video signs are distracting, around half believed they have a negative impact on

traffic safety, and 86% said there should be restrictions on video advertising (Smiley et

al., 2005).

2.4 Summary

In this chapter a detailed study of potential links between digital billboard and

traffic safety has been presented. The spatial and temporal analysis of crash rates at study

and control segments of a digital billboard have been mentioned as well.

Overall, literature reviews and crash analyses suggest that local conditions,

experimental settings, and other factors may play a role in the impact that driver

distraction due to advertising billboards has on traffic safety.

It should be also noted that existing research on the subject is limited due to a lack

of standardized methods and practices, data reliability, appropriate assumptions, relevant

hypotheses, and objective intentions. Consequently, new research on outdoor advertising

options and driver safety will prove paramount in the near future, especially because of

the dynamic state of the industry and the fact that many related studies are currently

outdated.

12

CHAPTER 3

SURVEY OF DRIVER‟S PERCEPTIONS

3.1 General

This chapter focuses on perceived impacts of digital advertising billboards on

driving performance of Alabama motorists from representative samples across the

lifespan. Perceived impacts were assessed through an online driver questionnaire survey

that documented perceptions and attitudes of drivers as they relate to roadside billboards.

The chapter describes the approach used to collect the data and summarizes findings from

drivers‟ responses.

3.2 Methods

One approach toward understanding transportation users‟ choices and behaviors is

through survey of drivers using questionnaires. In the present research, an online

questionnaire instrument was developed and used to gather and analyze data from

Alabama road user‟s perceptions and attitudes related to roadside advertising billboards

(SurveyMonkey). The questionnaire included a total of 22 questions that assessed several

variables of interest including demographic information (e.g., age, ethnicity, and gender),

exposure (driving patterns and experience, frequency of billboard encounters), driver‟s

behaviors, attitudes, and perceptions toward billboards with respect to safety and

efficiency, and respondents‟ stated preferences regarding placement, frequency and

regulation of roadway advertising billboard. To ensure random sampling, a company

specialized in web based surveys was hired to recruit a diverse group of survey

participants. In order to be eligible to participate in the survey, subjects had to possess a

valid driver‟s license and reside in Alabama.

13

3.3 Analysis

In aggregate, 295 respondents participated in this survey. Incomplete

questionnaire responses were omitted in order to maintain consistency for analysis.

Eventually, responses from 231 across the lifespan were used in the analysis. The

questionnaire extracted information related to driver demographics, driving experience

level, perception towards billboards, in general, and digital billboards, in particular,

attitudes related to use of information billboards, and perceptions on traffic safety with

respect to billboards and digital billboards. Participants‟ questionnaire responses were

collected and then processed using „Microsoft Excel‟ for further analysis.

3.4 Results

Out of 231 questionnaire respondents, 133 (57.58%) were male and 98 (42.32%)

were female drivers. Aggregate responses from the questionnaire are summarized in

Table 3.1.The findings reveal that 45.89% of respondents find billboards distracting in

general, and an overwhelming 67.53% perceive DBBs as more distracting than static

billboards.

14

Table 3.1: Aggregate Response from Online Questionnaire Survey

Question or Information Response % of total respondents

Are billboards distracting in general? Yes 45.89

No 31.60

Not sure 22.51

Do you think that DBBs are more

distracting than static billboards?

Strongly agree 22.08

Agree 45.45

Neither agree nor disagree 20.35

Disagree 11.26

Strongly disagree 0.87

Are you more likely to read a message

on a digital billboard than a static one?

Yes 48.92

No 38.10

Not sure 12.99

Do you glance long enough at a DBB to

read the entire message?

Rarely 25.54

Sometimes 42.86

Often 16.02

It depends on message 15.58

How often do you slow down to read a

DBB message?

Rarely 87.88

Sometimes 10.82

Often 1.30

How often do you use the information

from DBBs?

Rarely 74.46

Sometimes 23.81

All the time 1.73

Should there be restrictions on all

billboard locations for the purpose of

traffic safety?

Yes 61.90

No 16.02

Not sure 22.08

Should there be restrictions on the size

and number of digital billboards?

Yes 59.74

No 18.61

Not sure 21.65

Moreover, the majority responded that they are more likely to read a message on a

DBB rather than a static billboard. The majority (58.88%) also admitted that they stare at

a DBB long enough to read the entire message but they rarely slow down (87.88%) when

doing so. Interestingly, while responders admit that the messages posted on DBBs

15

capture their attention, three fourths of them (74.46%) state that they rarely use the

information.

For further analysis, the drivers were categorized into 7 age classes as

summarized in Figure 3.1. Approximately 13% of responders were under 20 years of age

and 11.26% were older than 55. The responses were then stratified according to the age

of the participants.

Figure 3.1: Number of Respondents with Age Class

When asked about their perception as it related to billboard distraction, 106

respondents (45.89%) reported that billboards cause „distraction.‟ The respondents in the

56-65 year old bracket had maximum rate of agreement on the issue of distraction from

presence of billboard (65%). The younger driver population, i.e., drivers of ≤20 years and

21-25 years of age also had a high percentage of agreement that the billboards cause

distraction (53.33% and 46.34%, respectively). The findings are summarized in Figure

3.2with the original survey question displayed at the top of the figure. The findings from

other survey questions will be represented in the same manner.

When asked if DBBs are more distracting than static billboards, nearly half of the

respondents (45.45%) agreed on the greater distracting power of the digital billboards.

16

Also, as shown in Figure 3.3, approximately 56% of those 21-25 years of age and 53% of

teen drivers ((≤20 years) agreed that digital billboards are more distracting than

traditional billboards.

Figure 3.2: Perception on Distraction by Billboards with Respect to Age

Figure 3.3: Perception on More Distraction Potential of Digital Billboard with Respect to

Age

Q. Are billboards distracting in general?

Q. Do you think digital billboards are more distracting than the regular (static) billboards?

17

So, it can be inferred that the rate of acceptance of potential distraction by digital

billboards in this study was higher among young drivers.

Almost half of the respondents also mentioned that they are more likely to read

messages from digital billboards (48.92%). This shows a clear intention of the road users

to be tempted by messages from digital billboards. Taking gender into consideration, the

tendency was greater among male drivers (52.63%) than their female counterparts

(43.88%). Interestingly, as depicted in Figure 3.4, this response was fairly consistent

across all age groups, including the elderly.

Figure 3.4: More Likeliness to Read Digital Billboard with Respect to Age

The analysis also revealed that over 42% of the road users sometimes glance at

the digital billboard for significantly long time. Although the exact time was not

described, the term „long‟ may be akin to several seconds. The scenario of long glance at

digital billboard was further broken down by age class and the results are shown in

Figure 3.5.

Q. Are you more likely to read message on a digital billboard than that on a static one?

18

Figure 3.5: Long Glance at Digital Billboard with Respect to Age

More than half (56.67%) of the young drivers (≤20 years of age) „sometimes‟

looked at the digital billboard for a long time, which is quite natural because the

respondents of this age might have a curiosity to the appearance and messages of digital

billboards. Though they sometimes glance for a long time, a small percent of drivers

across the life span reported doing it „often‟.

It can be deduced from the analysis of the responses that the overwhelming

majority of the questionnaire participants (87.88%) had a rare tendency to slow down

near digital billboards. Very small percentage of the drivers „sometimes‟ reduced their

speed (10.82%). Figure 3.6 shows the result of „slow down at digital billboard‟ scenario

based on age. The youngest driver group (≤20 years) rarely reduced their vehicle speed

disregarding the presence of digital billboard.

Q. Do you glance long enough at a digital billboard to read the entire message?

19

Figure 3.6: Slow Down to Digital Billboard with Respect to Age

Interestingly, most of the participants stated that they rarely used information

from digital billboards, and just over one-fifth of them (23.81%) used the information

sometimes. The rate was highest (36.84%) for participants between 46 and 55 years of

age. As can be seen in Figure 3.7, the youngest population group and the older population

(>65 years) showed almost no intention to use digital billboard‟s information.

The survey participants were also asked about their perception on the restriction

of locations of all billboards. The result is depicted in Figure 3.8.

The participants were also asked for their opinion on the restriction of size and

number of digital billboards. The result has been shown in Figure 3.9.

Q. How often do you slow down to read a digital billboard message?

20

Figure 3.7: Use of Information from Digital Billboard with Respect to Age

Figure 3.8: Perception on Restriction on Location of Billboards with Respect to Age

63.33%

53.66%

63.01% 61.90%

47.37%

80.00%83.33%

30.00%

17.07%12.33% 14.29%

21.05%

10.00%

0.00%

6.67%

29.27%24.66% 23.81%

31.58%

10.00%16.67%

0

10

20

30

40

50

60

70

80

90

≤20 years 21-25 years 26-35 years 36-45 years 46-55 years 56-65 years >65 years

Pe

rce

nta

ge o

f R

esp

on

se

Age

Age vs. Perception on Restrication on Location of Billboards

Yes

No

Not sure

Q. How often do you use information from digital billboards?

Q. Should there be restrictions on all billboard locations for the purpose of traffic safety?

21

Figure 3.9: Perception on Restriction on Size and Number of Digital Billboards with

Respect to Age

The above two questions have produced quite similar responses across all the age

groups which is an interesting finding. Most of the drivers surveyed think that there

should be stricter restrictions on location of all billboards and also on the size and number

of digital billboards for safety purpose.

Apart from the general analysis of the responses between genders and age groups,

chi-square tests have been performed across age groups and gender separately. The

observed values for the chi-square test have been found from the survey itself and the

expected values have been determined. The details of the chi-square test have been

presented in Appendix A (Tables A1 through A8). The results of the statistical analysis

are depicted in Tables 3.2 and 3.3.

The probability or p-values from Table 3.2 (in all cases greater than significance

level 0.05) suggest that there is no significant difference among responses across

different age groups of drivers when asked for their perceptions (e.g. if billboards are

66.67%

46.34%

64.38%

57.14%52.63%

75.00%

50.00%

26.67%24.39%

13.70%16.67%

21.05%

10.00%

33.33%

6.67%

29.27%

21.92%26.19% 26.32%

15.00% 16.67%

0

10

20

30

40

50

60

70

80

≤20 years 21-25 years 26-35 years 36-45 years 46-55 years 56-65 years >65 years

Pe

rce

nta

ge o

f R

esp

on

se

Age

Age vs. Perception on Restriction on Size & Number of DBBs

Yes

No

Not sure

Q. Should there be restrictions on the size and number of digital billboards for traffic safety?

22

distracting) and/or intended actions (e.g. slow down before digital billboard to read entire

message) to specific survey questions.

Similarly From Table 3.3 it can be implied that, there is no significant differences

between the responses of male and female drivers.

Table 3.2: Chi-square Test Result for Age Groups

Notion/Information/Query Degrees of

freedom (DF)

Chi-squire (χ2)* Probability

greater than Chi-

squire

(P>χ2) Are billboards distracting in general?

14 15.134 0.3691



28 16.886 0.9508


on a digital billboard than a static

one?

14 6.882 0.9392

Do you glance long enough at a DBB

to read the entire message?

21 18.591 0.6114

How often do you slow down to read

a DBB message?

14 13.018 0.5251


from DBBs?

14 15.309 0.3574



traffic safety?

14 16.396 0.2898

Should there be restrictions on the

size and number of digital billboards

14 17.101 0.2508

*Chi-square value derived from Pearson Chi-square test

From Table 3.2 it has been found that the probability (p-values) for all cases is

greater than the significance level (0.05). It means that the difference in responses across

age groups is not statistically significant. Similarly, from Table 3.3 it can be deduced that

there is no significant difference of the responses between male and female participants

when asked about their perception and/or potential actions (response of survey

questions). In other words, the responses are fairly consistent across the lifespan and

between sexes.

23

Table 3.3: Chi-square test Result for Male and Female

Notion/Information/Query Degrees of

freedom (DF)

Chi-squire

(χ2)* Probability greater than

Chi-squire (P>χ2) Are billboards distracting in general?

2 0.883 0.6431



4 2.409 0.6610


on a digital billboard than a static

one?

2 2.450 0.2938

Do you glance long enough at a DBB

to read the entire message?

3 3.348 0.3410

How often do you slow down to read

a DBB message?

2 0.782 0.6763


from DBBs?

2 2.154 0.3405



traffic safety?

2 4.763 0.0924

Should there be restrictions on the

size and number of digital billboards

2 3.232 0.1987

*Chi-square values derived from Pearson Chi-square test

3.5 Summary

The analysis of questionnaire surveys produced interesting insights regarding the

perceptions and attitudes of Alabama drivers with respect to digital advertising billboards.

Among other findings, road users perceived digital billboards as more dangerous than

their static counterparts and recommended stricter regulations of digital advertising

billboards. It has also been revealed that drivers do have sometimes or often a long glance

at digital advertising billboards, yet rarely slow down. This behavior might be a matter of

concern as it could increase the potential risk for traffic crash occurrence.

24

CHAPTER 4

CRASH RECORDS ANALYSIS: DATA COLLECTION CONSIDERATIONS

4.1 General

Data are mandatory for any type of research or analysis. The data type, quality

and quantity, and data collection method entirely depend on the purpose of the research

and economic feasibility associated with the study. A suitable data collection plan is

therefore of utmost importance to conduct a study. In this thesis, data were collected to

identify digital billboard locations, analyze crash data and find potential relationship

between crashes occurrence and presence of digital billboard.In the following subsections

data requirements, data collection methodology and other relevant features related to data

collection will be discussed.

4.2 Approach

The objective of this part of the study was to examine potential correlation

between presence of the digital billboards along the interstate routes of Alabama and

traffic safety. In doing so, historical crash records were retrieved and analyzed to allow

comparisons of crash rates in areas of potential influence of digital billboards to crash

rates in control segments downstream of digital billboard locations. This objective was

met in a series of steps that are depicted in Figure 4.1.

The digital billboards were identified using Google maps, digital advertising

company (Daktronics, and Lamar) websites, existing database, and other online

resources. Initially, a total of 26 digital billboards were identified along major interstate

freeway in the Birmingham, Mobile, Montgomery, and Huntsville regions. Three

25

billboards have been discarded as they were far away from the road (significant lateral

distance or offset).

Figure 4.1: Steps Associated with the Alabama Crash Analysis Study

After the identification of the digital advertising billboards locations, their

influence zone has been set. The influence zone (i.e., zone within which the driver might

be distracted by the digital advertising billboard) consists of two segments. The first

segment is upstream of the digital advertising billboard location (with respect to the

oncoming vehicle facing the digital face) and the second one is immediately beyond the

digital billboard. The former distance has been selected based on „visibility‟ of the drivers

in a clear, sunny day with no obstruction (another static or digital billboard, tree etc.) and

has been considered as 0.5 mile (with 0.1 mile accuracy). The concept of the second

segment has come from the fact that the drivers might continue to be mentally distracted

by the digital billboard for a short while after they passed the billboard location. This

distance has been chosen as a minimum 0.05 mile (with 0.02 mile accuracy). In some

cases the roadway curvature and other obstacles have restricted the visibility to 0.35

Identify Digital Billboards (DBBs)

Set Influence Zone of each DBB

Obtain & Analyze Crash Data

Seek Relationship between Crash and DBB

26

miles (driver cannot see the digital billboard beyond this distance while approaching the

billboard).

The “control site” for each digital billboard study location was a non-influence

zone represented by another segment located downstream from the digital billboard. The

length of this segment has been set at a minimum of 0.2 miles and cannot exceed the

corresponding upstream segment length. Figure 4.2 shows a typical study location. In this

study, the digital advertising billboard influence zone (study section) has been named as

the upstream segment (u/s) and the non-influence zone (control section) refers to the

downstream segment (d/s).

Figure 4.2: Typical Study Location

This step (i.e., study location identification) has resulted in the omission of nine

more digital billboard sites as they were very close to interchanges where the traffic

volume changes abruptly. Moreover, crashes associated with those sites only happened at

interchange or intersection locations, rather than the mainline. Two other digital

billboards could not meet the minimum downstream length criterion, and thus had to be

DBB

u/s

0.2 mile ≤ d/s ≤ u/s

(corresponding)

(0.05±0.02) mile

(0.5±0.1) mile

Travel direction

27

eliminated. In addition, two digital billboards had upstream and/or downstream segments

which contained static billboards and therefore could not be considered. One other digital

billboard was discarded as the billboard was situated at a curve section of the road that

affected driver visibility. Finally, one more site was eliminated since it had less than 2

crashes during the 5-year analysis period. Eventually, eight digital billboards were

selected for further analysis and those sites provided a good sample for the intended

analysis. The digital billboard locations for this study are depicted in Figure 4.3 on a

county-by-county basis and a brief description of the study locations characteristics is

presented in Table 4.1.

A total of five years (2008 to 2012) of crash data has been analyzed in this study.

The crash data has been gathered from „Critical Analysis Reporting Environment

(CARE)‟ website. The data includes crash ID (unique identifier), location (county and

city), year, month, week of the month, day of the week, time of the day, roadway

environment (rural, urban, suburban), highway classification (interstate, US highway

etc.), manner (type) of crash (e.g. single vehicle crash, rear end, side swipe etc.), crash

severity (fatal, possible injury, incapacitating injury etc.), milepost, route name (e.g.

interstate 65, interstate 459, highway 31 etc.), longitude, latitude and a lot other different

information. At first, the data have been filtered on the basis of the year. Then further

sorting has been done to isolate the interstate crashes only. ArcGIS 10.1 has been used to

plot the crashes on the map. The billboard locations which have been already marked in

the map are then superimposed with the crash locations (latitude and longitude). This

operation has enabled to count the exact number of crashes at the billboard influence

zone (u/s) and non-influence zone (d/s). The average annual daily traffic (AADT) data

28

has been obtained from ALDOT records and used for the crash analysis in order to

determine crashes per million vehicle miles per year to make the analysis more logical.

Figure 4.3: Spatial Representation of Study Locations (On County-by-County Basis)

29

Table 4.1: List of Alabama Study (u/s) and Control (d/s) Sections

Lo

cati

on

ID

Cit

y

Co

un

ty

Ro

ute

Dir

ect

ion

of

Tra

vel

Ro

ad

Sid

e

La

nd

Use

MP

u/s

Seg

men

t

Len

gth

, L

(mil

es)1

d/s

Seg

men

t

Len

gth

, L

(mil

es)2

1 Mobile Mobile I-65 SB R Urban 7.31 0.453 0.453

2 Mobile Mobile I-65 NB R Urban 5.01 0.467 0.237

3 Mont-

gomery

Mont-

gomery

I-85 SW (WB) R Suburban 10.07 0.396 0.396

4 Madison Madison I-565 NE (EB) R Urban 10.78 0.373 0.373

5 Huntsville Madison I-565 NE (EB) R Urban 14.87 0.353 0.353

6 Huntsville Madison I-565 SW (WB) R Urban 14.87 0.486 0.207

7 Bessemer Jefferson I-459 NW (WB) R Urban 16.56 0.505 0.505

8 Bessemer Jefferson I-20/59 SB R Suburban 113.46 0.497 0.497

1Upstream length includes 0.05 (±0.02) miles downstream of digital billboard

2Downstream length equals to or less than corresponding upstream length

Figures 4.4 and 4.5 show snapshot of two study locations (location 7 and 8).

Figure 4.4: Location ID 7 on I-459 in Bessemer (Jefferson County)

30

Figure 4.5: Location ID 8 on I-20/59 in Bessemer (Jefferson County)

4.3 Summary

In this chapter the whole approach for selection of study and control sections has

been discussed. Although a significant number of digital billboards are situated near the

interstate routes, a large portion of the locations could not be considered as they did not

satisfy the selection criteria. In selecting sites for comparison, consistency in length and

characteristics of study and control segments has been maintained to the highest level

possible.

31

CHAPTER 5

CRASH DATA ANALYSIS: METHODOLOGY AND RESULTS

5.1 General

This chapter focuses on the assessment of the traffic safety impact of digital

advertising billboards in the State of Alabama. Following the site selection and data

gathering approach detailed in Chapter 4, the next paragraphs summarize concepts and

results from the crash data analysis at the selected study sites.

5.2 Data Analysis Procedure

First, analysis of crash trends was performed to gain a better understanding of

crash trends at the study sites over a 5 year span (2008 through 2012). Then crash rates

per million vehicle miles traveled at the DBB influence areas (u/s) and non-influence

areas (d/s) were determined and comparisons were made to establish if there exists any

relationship between presence of digital billboard and frequency of crash occurrence.

In doing so, the vehicle miles of travel (VMT) for each year (year 2008 through

2012) were calculated using the following equation:

Vehicle miles travel (VMT) at any year i = AADT of year i * 0.5 * L* 365…………. (1)

where,

AADT = Average annual daily traffic (both direction) at billboard influence zone in

vehicles/day, and

L = Length of billboard influence zone in miles.

AADT is actually the daily traffic volume collected from the traffic counts data of

exactly one year and then divided by 365 days to find the daily volume (on average). This

AADT comprises of vehicle counts for both directions of road. But the distraction (and

32

perhaps resulting crash) is directional as the digital face of billboard is supposed to

convey message for a particular travel direction (unless both faces are digital). Therefore

the AADT has been multiplied by 0.5 to convert it to one directional volume. The symbol

„L‟ refers to the length of the billboard influence zone as defined in Figure 2. As the

VMT for one year is considered, the one directional volume (for one day) has been

multiplied by 365 days.

Afterwards, the crash rates (crashes per million vehicle miles) for all the study

locations from years 2008 to 2012 are shown in Table 4.2. The average annual crash rate

(CR) has been calculated using Equation (2).

CRj = [(Ncj * 106)/N] / [(VMTj,total)/N] ……………………………………………….. (2)

where,

CRj = Average annual crash rate for location j (in crashes per million vehicle miles)

Ncj = Total number of crashes (from 2008-2012) at location j in direction of digital face

(one direction)

N = Crash data analysis period (in years) = 5

VMTj,total = Total vehicle miles traveled in direction of digital face (one direction) at

location j = (VMTjin 2008 + VMTj in 2009 + VMTj in 2010 + VMTj in 2011 +

VMTj in 2012)

The crash rates have been determined for both the influence (upstream) and non-

influence (downstream) zones of digital billboards. The crashes have been counted based

on the direction of the vehicles approaching the digital face of the billboard (upstream)

and the vehicles passed the digital face (downstream).

33

So the combined VMT (of 5 years) has been used to calculate average annual

crash rates in each location. The number of crashes at each year for a particular location

was small and therefore total number of crashes for five years was used in determination

of the crash rate.

5.3 Results

5.3.1 Analysis of Crash Records Trends

Crash frequencies (i.e., number of crashes) for the 5 year study period at the study

locations were plotted to observe variations by a. year, b. month of the year, c. day of the

week, and d. time of the day. The details of the crash records trends are summarized in

Appendix B (Tables B1 through B4).

Figure 5.1: Aggregate Crash Frequency by Year

Crash frequency by year: Figure 5.1 shows the variation of aggregate crash

frequency for the years 2008 to 2012. It can be seen that the number of crashes has been

decreasing gradually since 2009.

0

2

4

6

8

10

12

14

16

18

20

Year 2008 Year 2009 Year 2010 Year 2011 Year 2012

Cra

sh F

req

ue

ncy

Frequency of All Study Crashes by Year

34

A comparison of crashes occurring in the DBB influence zone (u/s) and non-

influence zone (d/s) over the study period is shown in Figure 5.2. It can be seen that in

each and every year the number of crashes at DBB influence zones (u/s) surpassed the

number of crashes at control (d/s) segments.

Figure 5.2: Crash Frequency at DBB Influence Zones and Control Segments by Year

Crash frequency by month of the year: Figure 5.3 shows the variation of crash

frequency at all study sites combined by month over the study period (2008 through

2012). The figure suggests that the digital billboard locations experienced the maximum

number of crashes in the winter months, with the peak taking place in November.

Comparison of crash frequencies at DBB influence (u/s) and non-influence zones

(d/s) shows mixed results and no specific pattern of crash frequency can be determined.

Both influence (u/s) and non-influence zones (d/s) show higher numbers of crashes

during the winter months.

0

2

4

6

8

10

12

14

Year 2008 Year 2009 Year 2010 Year 2011 Year 2012

Cra

sh F

req

ue

ncy

Frequency of Crashes by Year

Crashes as DBB influence zone (u/s)

Crashes at control segment (d/s)

35

Figure 5.3: Crash Frequency at Study Sites by Month

Crash frequency by day of the week: Figure 5.4 shows the variation of crash

frequency at all study sites combined by day of the week over the study period.

According to the data, the maximum number of crashes occurred on Sunday. Comparison

of the number of study crashes at the DBB influence and non-influence zone does not

suggest any specific trends.

Figure 5.4: Crash Frequency at Study Sites by Day of the Week

Crash frequency by time of the day: Figure 5.5 shows the variation of crash

frequencies at all study sites combined by time of the day over the 5 year study period.

0

2

4

6

8

10

12

14

16

Cra

sh F

req

ue

ncy

Frequency of Crashes by Month


Crashes at DBB influence zone (u/s)

0

2

4

6

8

10

12

14

16

Monday Tuesday Wednesday Thursday Friday Saturday Sunday

Cra

sh F

req

ue

ncy

Frequency of Crashes by Day of the Week



36

The highest number of crashes occurred at 5:00AM to 5:59AM time period followed by

periods coinciding with morning, lunch, and afternoon peak periods.

Figure 5.5: Crash frequency at Study Sites by Time of the Day

5.3.2 Crash Analysis Results

Crash summary by location and paired t-Test for significance: Table 5.1 shows

the summary statistics of crash rates at the eight study sites (both for the DBB influence

and non-influence zones). As far as the number of crashes is concerned, the majority of

the sites experienced more crashes in the DBB influence zone than the control

(downstream non-influence zone). Over the analysis period, a total of 49 crashes took

place at all study DBB influence zones (u/s) combined as opposed to on 28 in the DBB

non-influence zones (d/s). Two locations (locations 6 and 8) reported 3 and 9 crashes

respectively in the DBB influence zone and none in the non-influence zone, hinting a

potential influence from the DBB presence.

The data analysis further revealed that crash rates at DBB influence zones were

higher at some of the study locations (namely locations 3, 4, 6, 8) but lower in the

0123456789

10

12

:00

Mid

nig

ht

to 1

2:5

9 …

1:0

0 A

M t

o 1

:59

AM

2:0

0 A

M t

o 2

:59

AM

3:0

0 A

M t

o 3

:59

AM

4:0

0 A

M t

o 4

:59

AM

5:0

0 A

M t

o 5

:59

AM

6:0

0 A

M t

o 6

:59

AM

7:0

0 A

M t

o 7

:59

AM

8:0

0 A

M t

o 8

:59

AM

9:0

0 A

M t

o 9

:59

AM

10

:00

AM

to

10

:59

AM

11

:00

AM

to

11

:59

AM

12

:00

No

on

to

12

:59

PM

1:0

0 P

M t

o 1

:59

PM

2:0

0 P

M t

o 2

:59

PM

3:0

0 P

M t

o 3

:59

PM

4:0

0 P

M t

o 4

:59

PM

5:0

0 P

M t

o 5

:59

PM

6:0

0 P

M t

o 6

:59

PM

7:0

0 P

M t

o 7

:59

PM

8:0

0 P

M t

o 8

:59

PM

9:0

0 P

M t

o 9

:59

PM

10

:00

PM

to

10

:59

PM

11

:00

PM

to

11

:59

PM

Un

kno

wn

Cra

sh F

req

ue

ncy

Frequency of Crashes by Time of the Day



37

remaining ones. When considering all sites combined, the crash rates at DBB influence

zones were 29% higher than those of their counterparts representing non-influence zones,

indicating a higher likelihood for crash occurrence in the presence of a digital billboard.

Table 5.1: Crash Summary Statistics at the Digital Billboard Locations (Aggregate

Value)

Location City

Total

VMT

DBB Influence

Zone (u/s)

DBB Non-Influence

Zone (u/s)

Percent

Change

inCrash

Rate

Len.

(mi)

Total

Crash

Count

Crash

Rate*

Len.

(mi)

Total

Crash

Count

Crash

Rate*

1

Mobile

30505326

0.453

6

0.197

0.453

7

0.229

16.67

2

Mobile 40099539 0.467 15 0.374 0.237 9 0.442 18.23

3

Montgomery 16523813 0.396 5 0.303 0.396 2 0.121 -60.00

4

Madison 19848580 0.373 4 0.202 0.373 1 0.050 -75.00

5

Huntsville 29193700 0.353 3 0.103 0.353 4 0.137 33.33

6

Huntsville 40193026 0.486 3 0.075 0.207 0 0.000 -100.00

7

Bessemer 23026801 0.505 4 0.174 0.505 5 0.217 25.00

8

Bessemer 22537757 0.497 9 0.399 0.497 0 0.000 -100.00

Total crashes 221928541 3.53 49 0.221 3.021 28 0.156 -29.19 *Crash rate refers to „average annual crash rate‟ and is in crashes per million vehicle miles per year

A paired t-test was performed to test whether the presence of DBB has a

significant impact on crash occurrence. The null hypothesis was set as μD=0 indicating

that the means of crash counts at the two zones (i.e., u/s and d/s) are the same. For the

level of significance of α=0.05, the criterion was to reject the null hypothesis if t >1.415

(d.f.=7) where,

𝑡 = 𝐷−0

S D

𝑛

........................................................................................................................ (3)

and D and SD are the mean and standard deviation of the differences (D=2.625 and

38

SD=3.623) and n=8. It was found that t=2.05>1.415, thus, the null hypothesis must be

rejected at level of significance α=0.05.

In the following subsections, crashes have been summarized by type and severity.

The details of the crashes by location, ID, type and severity are shown in appendix B

(Table B5).

Summary by crash type: The summary statistics of the crash type for all the eight

study sites are shown in Table 5.2. It can be seen that the study locations experienced a

total of six types of specified crashes. There is another category which does not define the

types of crashes precisely (e.g. record from paper system).

Table 5.2: Summary Statistics by Crash Type

Crash Type Upstream (u/s) Downstream (d/s) Percent Change

in Crash Rate Crash Count

Crash Rate1

Crash Count Crash Rate1

Non-collision

1 0.005 0 0 -100.00

Single Vehicle

Crash

7 0.032 8 0.045 40.63

Angle (front to

side) Same

Direction

1 0.005 0 0 -100.00

Rear End

11 0.050 7 0.039 -22.00

Side Impact (90

degrees)

1 0.005 0 0 -100.00

Sideswipe –

Same Direction

6 0.027 0 0 -100.00

Record from

Paper System

22 0.099 13 0.072 -27.27

Total Crashes 49 0.221 28 0.156 -29.19 1Crash rate refers to „average annual crash rate‟ and is in crashes per million vehicle miles per year

Among the definite crash types, the sideswipe and rear end crashes are clearly

overrepresented at the DBB influence areas. In fact, non-collision, angle (front side; same

39

direction), side impact (90 degrees) and sideswipe (same direction) type crashes occurred

only at the DBB influence zones.

Summary by crash injury severity: Table 5.3 shows the severity of crashes at the

DBB influence- and non-influence zones in aggregate for all study locations.

Table 5.3: Summary Statistics by Crash Injury Severity

Crash

Severity

Upstream (u/s) Downstream (d/s) Percent Change

in Crash Rate Crash Count Crash Rate1

Crash Count Crash Rate1

Fatal Injury

2 0.009 1 0.006 -33.33

Incapacitating

Injury

6 0.027 1 0.006 -77.78

Non-

incapacitating

Injury

0 0 2 0.011 ---

Possible Injury

4 0.018 1 0.006 -66.67

Property

Damage Only

(PDO)

35 0.158 22 0.123 -22.15

Unknown

2 0.009 1 0.006 -33.33

Total Crashes 49 0.221 28 0.156 -29.19 1Crash rate refers to „average annual crash rate‟ and is in crashes per million vehicle miles per year

There are a total of five levels of specific crash severity mentioned here (unknown

is not specific class). A total of three fatalities (two along I-65 in Mobile in 2011 and

2008, one along I-565 at Huntsville in 2009) have been found, two of which occurred at

DBB areas of influence. It should be noted that the number of crashes is very small and

does not allow for in depth statistical analysis. Still, the data show that a higher number

of more severe crashes occur at DBB influence zones, compared to non-influence zones,

once again suggesting a link between distraction from DBB presence and crash severity.

40

5.4 Discussion

The analysis on crash rates by location has revealed an overall lower crash rate at

the DBB non-influence zone compared to the DBB influence zone (29% lower).

Individual site data show that crash rate has been decreased specifically at four locations.

The statistical analysis further showed a statistically significant difference in the

frequency of crashes reported at the DBB sites when compared to the control study sites.

This finding indicates an association between DBB presence and crash occurrence at the

Alabama case study.

The analysis of the crashes based on crash type revealed that the sideswipe and

rear end crashes (often related to driver distraction) were clearly overrepresented at the

DBB influence areas. Furthermore, consideration of crash severity showed evidence of

overrepresentation of certain types of crashes at DBB influence zones (especially fatal,

incapacitating injury, and possible injury crashes). However, the sample size is small to

allow for a detailed statistical analysis.

5.5 Summary

This chapter focused on the analysis of crash data at digital billboard study sites

and control sections. The summary of the analysis has been presented in terms of crash

rates, type and severity. The comparison of crash rates at the digital billboard study

sections and control sections suggests there is statistically significant difference between

crash rates. In other words this finding is indicative of a relationship between crash

occurrence and presence of digital advertising billboards.

41

CHAPTER 6

CONCLUSIONS AND RECOMMENDATIONS

6.1 Summary of Research

This research has aimed to investigate the potential relationship between the

presence of digital advertising billboards and occurrence of crashes at interstates in

Alabama. As a preliminary study, an online driver questionnaire survey has been

conducted to gain information about the drivers‟ behaviors, attitudes and perceptions

toward digital billboards. The questionnaire survey has also included demographic

information and experience of the drivers. After obtaining the initial survey result, the

impact of digital billboard on traffic safety (in terms of driver distraction) on high-speed,

limited-access facilities was explored at eight study sites in Alabama. The methodology

of crash investigation in both states relied on comparing the crash rate statistics upstream

and downstream each billboard location. The upstream and downstream segments at each

billboard location were selected so that they experienced the same traffic and geometric

conditions, i.e., number of lanes, roadside features, no weaving maneuvers, etc. Total 77

crashes were used in the analysis. The overall result suggested that the increase 29%

although the site specific change was varying.

6.2 Implications for Practice

Although the crash analysis in Alabama consistently revealed that the impact of

digital billboards on traffic safety is small, there is still a correlation between driver

distraction and traffic safety (though small). Digital billboard manufacturers should

design these billboards with the minimal amount of animations to minimize the impact of

distraction on drivers. It is also recommended to avoid installing digital billboards on

42

sections with horizontal and vertical alignments and locations with high historical

number of crashes to significantly diminish the impact of driver distraction.

Among other findings, the survey of users highlights the need for better regulation

of digital advertising billboards in the future. The study recommends reevaluation of

current legislation and regulation for controlling outdoor advertising both at the federal

and state level. Updates of regulations shall consider restrictions in the frequency,

placement and operation of digital advertising billboards in order to protect the safety of

the public and reduce unnecessary cluttering and visual pollution.

6.3 Limitations and Future Research

It should be noted that the findings from the crash analysis in the state of Alabama

was based on relatively small sample of locations and relatively small segment lengths. It

is recommended to validate the results of the crash analysis using larger sample sizes and

longer segments. Future research could compare the findings of the crash analysis in

Alabama with other states to determine how the impact of digital billboard on traffic

safety varies across states. Crash analysis on other roadway facilities that carry digital

advertising billboards, e.g., arterials can be also conducted to evaluate the potential safety

impacts on DBB in such settings. It will be quite a challenge to do the same research for

arterials as uniform flow conditions, large segments of straight roadway and adequate

visibility could not be found easily as in case of Interstate routes. It is almost impossible

to carry on similar research on urban streets due to interrupted flow conditions (presence

of traffic signals at intersections). The findings from this research can be compared with

the driving simulator studies. It can be checked if the types of billboards (e.g., food

advertisements vs. public health announcements vs. variable message signs) evoke

43

significantly more driver distraction than others or whether billboard placement (i.e.,

right vs. left) has a differential impact. Studies could also consider what specific aspects

of billboards (e.g. graphics, slogans, and exit numbers) divert drivers‟ attention from the

roadway more readily. Moreover, there may be a study concerning the influence of on

premises vs. off premises billboard on driver‟s driving behavior. Empirical Bayes

Method (EBM) can also be conducted as discussed in the literature review. In order to be

effective, crash modeling can also be done in future.

44

LIST OF REFERENCES

Smiley, A., Persaud, B., Bahar, G., Mollett, C., Lyon, C., Smahel, T., &Kelman, W. L.

(2005), Traffic Safety Evaluation of Video Advertising Signs, Transportation Research

Record: Journal of the Transportation Research Board, No. 1937, Washington, D.C., pp.

105-112.

Tantala, A. M., &Tantala, M. W. (2007), A Study of the Relationship between Digital

Billboards and Traffic Safety in Cuyahoga County, Ohio. Submitted to the Foundation for

Outdoor Advertising Research and Education (FOARE), 1850 M Street, NW, Suite 1040,

Washington, DC 20036-5821.

Henson, S.C. (2009), Digital Billboard Safety among Motorists in Los Angeles,

Department of Geography, California State University, Northridge, Los Angeles, CA.

Tantala, M. W., &Tantala, A. M. (2010a), A Study of the Relationship between Digital

Billboards and Traffic Safety in Albuquerque, NM. Report for the Foundation for Outdoor

Advertising Research and Education (FOARE), Washington, DC.

Tantala, M. W., &Tantala, A. M. (2010b), A Study of the Relationship between Digital

Billboards and Traffic Safety in Henrico County and Richmond, Virginia. Report for the

Foundation for Outdoor Advertising Research and Education (FOARE), Washington,

DC.

Outdoor Advertising Association of America [OAAA] (2013), Out of Home Media

Formats: Number of Out of Home Displays (2013), October 14, 2013, from

http://www.oaaa.org/OutofHomeAdvertising/OOHMediaFormats/OOHMediaFormats.as

px

Copeland, L. (2010), More Communities Banning „Television on a Stick.‟ USA Today.

Sisiopiku, V.P., Hester, D., Gan, A., Stavrinos, D., & Sullivan, A. (2013), Roadside

Advertising and Traffic Safety. Proceedings of the 3rdAnnual International Conference on

Civil Engineering, Athens, Greece.

Lee S., McElheny, M., & Gibbons, R. (2007), Driving Performance and Digital

Billboards. Report for Foundation for Outdoor Advertising Research and Education by

the Virginia Tech Transportation Institute (VTTI), Center for Automotive Safety

Research.

http://www.oaaa.org/OutofHomeAdvertising/OOHMediaFormats/OOHMediaFormats.aspx

http://www.oaaa.org/OutofHomeAdvertising/OOHMediaFormats/OOHMediaFormats.aspx

45

Institute for Road Safety Research, Roadside Advertising and Information (2013),

Available at: http://www.swov.nl/rapport/Factsheets/UK/FS_Advertising.pdf

Wallace, B. (2003), Driver Distraction by Advertising: Genuine Risk or Urban

Myth?Proceedings of the Institution of Civil Engineers: Municipal Engineer, Vol. 156,

No. 3, pp. 185-190.

Coetzee, J. (2003), The Evaluation of Content on Outdoor Advertisements. Presented at

the Southern African Transport Conference.

Farbry, J., Wochinger, K., Shafer, T., Owens N., Nedzesky, A. (2001), Research Review

of Potential Safety Effects of Electronic Billboards on Driver Attention andDistraction.

Publication FHWA-RD-01-071. FHWA, Office of Safety Research and Development.

Molino, J.,Wachtel, J.,Farbry, J., Hermosillo, M., Granda, T. (2009), The Effects of

Commercial Electronic Variable Message Signs (CEVMS) on Driver Attention and

Distraction: An Update. Publication FHWA-HRT-09-018, FHWA.

Wachtel, J. (2009), Safety Impacts of the Emerging Digital Display Technology for

Outdoor Advertising Signs. Project 20-7 (256), Final Report,NCHRP.

Crawford, R. (2010), Inside the Wachtel 2009 Digital Display Report: A Commonsense

Guide. United States Sign Council, Version 3.17.10.

Wachtel, J. (2011), Digital Billboards, Distracted Drivers. Planning, Vol. 77, Issue 3, pp.

25-27.

Online Driver Questionnaire Survey for Alabama (2014), SurveyMonkey, from

https://www.surveymonkey.com/r/?sm=l3OugwbHi4B3uQMdLq6vwhPni6oi7OWO0Cb

er7pnBSQ%3d

CARE 9.0.0 (2007), Critical Analysis Reporting Environment, User Manual, CARE

Research and Development Laboratory. Computer Science Department, The University

of Alabama (UA), available at caps.ua.edu/care.aspx.

http://www.swov.nl/rapport/Factsheets/UK/FS_Advertising.pdf

https://www.surveymonkey.com/r/?sm=l3OugwbHi4B3uQMdLq6vwhPni6oi7OWO0Cber7pnBSQ%3d

https://www.surveymonkey.com/r/?sm=l3OugwbHi4B3uQMdLq6vwhPni6oi7OWO0Cber7pnBSQ%3d

46

APPENDIX A

CHI-SQUARE TEST RESULT FROM ONLINE QUESTIONNAIRE SURVEY

47

Table A1: General Perception of Distraction by Billboard

Table A2: Use Information from Digital Billboard (DBB)

Table A3: More Likeliness to Read Messages from DBBs

Observed value

≤20 years (number) 21-25 years (number) 26-35 years (number) 36-45 years (number) 46-55 years (number) 56-65 years (number) >65 years (number)

16 19 36 12 9 13 1

10 13 20 15 7 5 3

4 9 17 15 3 2 2

30 41 73 42 19 20 6

Expected value


13.7662 18.8139 33.4978 19.2727 8.7186 9.1775 2.7532

9.4805 12.9567 23.0693 13.2727 6.0043 6.3203 1.8961

6.7532 9.2294 16.4329 9.4545 4.2771 4.5022 1.3506

30 41 73 42 19 20 6

p value 0.2855

Not significant

Observed value


25 28 62 26 12 14 5

5 12 9 15 7 6 1

0 1 2 1 0 0 0

30 41 73 42 19 20 6

Expected value


22.33766234 30.52813853 54.35497835 31.27272727 14.14718615 14.89177489 4.467532468

7.142857143 9.761904762 17.38095238 10 4.523809524 4.761904762 1.428571429

0.519480519 0.70995671 1.264069264 0.727272727 0.329004329 0.346320346 0.103896104

30 41 73 42 19 20 6

p value 0.2768

Not significant

Observed value


15 23 32 20 10 10 3

12 12 30 15 8 9 2

3 6 11 7 1 1 1

30 41 73 42 19 20 6

Expected value


14.67532468 20.05627706 35.70995671 20.54545455 9.294372294 9.783549784 2.935064935

11.42857143 15.61904762 27.80952381 16 7.238095238 7.619047619 2.285714286

3.896103896 5.324675325 9.480519481 5.454545455 2.467532468 2.597402597 0.779220779

30 41 73 42 19 20 6

p value 0.9488

Not significant

48

Table A4: Long Glance at Digital Billboard (DBB)

Table A5: Slowing Down to DBB

Table A6: Perception on More Potential Distraction Caused by DBB

Observed value


5 9 18 10 6 9 2

17 15 35 18 7 5 2

4 8 9 6 4 4 2

4 9 11 8 2 2 0

30 41 73 42 19 20 6

Expected value


7.662337662 10.47186147 18.64502165 10.72727273 4.852813853 5.108225108 1.532467532

12.85714286 17.57142857 31.28571429 18 8.142857143 8.571428571 2.571428571

4.805194805 6.567099567 11.69264069 6.727272727 3.043290043 3.203463203 0.961038961

4.675324675 6.38961039 11.37662338 6.545454545 2.961038961 3.116883117 0.935064935

30 41 73 42 19 20 6

p value 0.7010

Not significant

Observed value


30 35 64 35 17 18 4

0 5 9 6 1 2 2

0 1 0 1 1 0 0

30 41 73 42 19 20 6

Expected value


26.36363636 36.03030303 64.15151515 36.90909091 16.6969697 17.57575758 5.272727273

3.246753247 4.437229437 7.9004329 4.545454545 2.056277056 2.164502165 0.649350649

0.38961039 0.532467532 0.948051948 0.545454545 0.246753247 0.25974026 0.077922078

30 41 73 42 19 20 6

p value 0.3677

Not significant

Observed value


5 8 17 8 5 6 2

16 23 31 20 5 8 2

6 4 16 10 7 3 1

3 6 8 3 2 3 1

0 0 1 1 0 0 0

30 41 73 42 19 20 6

Expected value


6.623376623 9.051948052 16.11688312 9.272727273 4.194805195 4.415584416 1.324675325

13.63636364 18.63636364 33.18181818 19.09090909 8.636363636 9.090909091 2.727272727

6.103896104 8.341991342 14.85281385 8.545454545 3.865800866 4.069264069 1.220779221

3.376623377 4.614718615 8.216450216 4.727272727 2.138528139 2.251082251 0.675324675

0.25974026 0.354978355 0.632034632 0.363636364 0.164502165 0.173160173 0.051948052

30 41 73 42 19 20 6

p value 0.9323

not significant

49

Table A7: Restriction on Billboard Location

Table A8: Restriction on Number and Size of DBB

Observed value


19 22 46 26 9 16 5

9 7 9 6 4 2 0

2 12 18 10 6 2 1

30 41 73 42 19 20 6

Expected value


18.57142857 25.38095238 45.19047619 26 11.76190476 12.38095238 3.714285714

4.805194805 6.567099567 11.69264069 6.727272727 3.043290043 3.203463203 0.961038961

6.623376623 9.051948052 16.11688312 9.272727273 4.194805195 4.415584416 1.324675325

30 41 73 42 19 20 6

p value 0.2217

not significant

Observed value


20 19 47 24 10 15 3

8 10 10 7 4 2 2

2 12 16 11 5 3 1

30 41 73 42 19 20 6

Expected value


17.92207792 24.49350649 43.61038961 25.09090909 11.35064935 11.94805195 3.584415584

5.584415584 7.632034632 13.58874459 7.818181818 3.536796537 3.722943723 1.116883117

6.493506494 8.874458874 15.8008658 9.090909091 4.112554113 4.329004329 1.298701299

30 41 73 42 19 20 6

p value 0.4087

not significant

50

APPENDIX B

AGGREGATE CRASH ANALYSIS

51

Table B1: Crash by Year

Frequency of

Crashes

Cumulative

Frequency of Crashes

Percentage of Crashes Cumulative

Frequency of

Crashes Year 2008 123992 123992 19.55 19.55

Year 2009 123975 247967 19.54 39.09

Year 2010 129529 377496 20.42 59.51

Year 2011 128501 505997 20.26 79.77

Year 2012 128318 634315 20.23 100.00

Table B2: Crash by Month

Frequency of

Crashes

Cum. Frequency Percentage of

Crash

Cum. Percentage

of Crash January 50376 50376 7.94 7.94

February 50309 100685 7.93 15.87

March 55055 155740 8.68 24.55

April 53675 209415 8.46 33.01

May 54135 263550 8.53 41.55

June 49867 313417 7.86 49.41

July 50302 363719 7.93 57.34

August 52675 416394 8.3 65.64

September 51516 467910 8.12 73.77

October 55514 523424 8.75 82.52

November 54386 577810 8.57 91.09

December 56505 634315 8.91 100

Table B3: Crash by Day of the Week

Frequency of

Crash

Cum.

Frequency

Percentage of

Crash

Cum. Percentage

Sunday 60631 60631 9.56 9.56

Monday 92156 152787 14.53 24.09

Tuesday 95222 248009 15.01 39.1

Wednesday 93992 342001 14.82 53.92

Thursday 96800 438801 15.26 69.18

Friday 113775 552576 17.94 87.11

Saturday 81739 634315 12.89 100

52

Table B4: Crash by Time of the Day

Frequency of

Crashes

Cum. Frequency

of crashes

Percentage of

Crashes

Cum. Percentage

of Crashes 12:00 Midnight to 12:59

AM

8347 8347 1.32 1.32

1:00 AM to 1:59 AM 7232 15579 1.14 2.46

2:00 AM to 2:59 AM 6966 22545 1.1 3.55

3:00 AM to 3:59 AM 5966 28511 0.94 4.49

4:00 AM to 4:59 AM 5941 34452 0.94 5.43

5:00 AM to 5:59 AM 9260 43712 1.46 6.89

6:00 AM to 6:59 AM 15047 58759 2.37 9.26

7:00 AM to 7:59 AM 39602 98361 6.24 15.51

8:00 AM to 8:59 AM 27709 126070 4.37 19.87

9:00 AM to 9:59 AM 24779 150849 3.91 23.78

10:00 AM to 10:59 AM 28648 179497 4.52 28.3

11:00 AM to 11:59 AM 35461 214958 5.59 33.89

12:00 Noon to 12:59 PM 42521 257479 6.7 40.59

1:00 PM to 1:59 PM 41665 299144 6.57 47.16

2:00 PM to 2:59 PM 44370 343514 6.99 54.16

3:00 PM to 3:59 PM 57831 401345 9.12 63.27

4:00 PM to 4:59 PM 52837 454182 8.33 71.6

5:00 PM to 5:59 PM 55867 510049 8.81 80.41

6:00 PM to 6:59 PM 35953 546002 5.67 86.08

7:00 PM to 7:59 PM 24461 570463 3.86 89.93

8:00 PM to 8:59 PM 21249 591712 3.35 93.28

9:00 PM to 9:59 PM 17565 609277 2.77 96.05

10:00 PM to 10:59 PM 13776 623053 2.17 98.22

11:00 PM to 11:59 PM 10878 633931 1.71 99.94

Unknown 384 634315 0.06 100

53

Table B5: Crash ID, Type and Severity at DBB Locations (Aggregate Data)

Upstream Downstream

Crash ID Crash Type Crash

Severity

Crash ID Crash Type Crash

Severity DBB Location 1

Year 2012 2630619 Rear End (Front

To Rear)

PDO

2605099 Single Vehicle

Crash

Unknown

Year 2011

1609383 Side impact

(90 degrees)

PDO 1600855 Single Vehicle

Crash

PDO


Crash

PDO

Year 2010

666891 Rear End (Front

To Rear)

PDO

Year 2009

9001420 RecfrPapSys PDO 9677241 Single Vehicle

Crash

Non-

incapacitating

injury

9050750 RecfrPapSys PDO

9054666 RecfrPapSys Incapacitating

Injury

Year 2008

8525997 RecfrPapSys Fatal Injury 8034086 RecfrPapSys PDO

8067149 RecfrPapSys Possible

Injury

DBB Location 2

Year 2012


To Rear)

Possible

Injury


To Rear)

PDO


Crash

PDO 2654145 Rear End (Front

To Rear)

Possible

Injury

Year 2011


To Rear)

Unknown 1612858 Single Vehicle

Crash

Fatal Injury


To Rear)


Crash

PDO


To Rear)

PDO

Year 2010


To Rear)


To Rear)

Non-

incapacitating

injury


To Rear)


To Rear)

PDO


Crash

Incapacitating

Injury

Year 2009

9691334 Sideswipe

(Same direc.)


Crash

PDO


To Rear)

Incapacitating

Injury


9510281 RecfrPapSys PDO 9504954 RecfrPapSys PDO

Year 2008




Injury

54


DBB Location 3

Year 2012

Year 2011 1702297 Angle (front to

side) Same Direc.

PDO

1708228 Sideswipe

(Same direc.)

PDO

Year 2010

Year 2009 9698838 Sideswipe

(Same direc.)

PDO 9012357 RecfrPapSys PDO

Year 2008 8071492 RecfrPapSys PDO 8017865 RecfrPapSys PDO


DBB Location 4

Year 2012


Crash

PDO


To Rear)

PDO

Year 2010 672231 Non-collision PDO

699577 Sideswipe

(Same direc.)

PDO

Year 2009 9502339 RecfrPapSys Possible

Injury

Year 2008

DBB Location 5

Year 2012

Year 2011 1622752 Single Vehicle

Crash

Unknown


To Rear)

PDO 616 RecfrPapSys PDO


To Rear)

PDO

Year 2009 9514769 RecfrPapSys Fatal Injury


DBB Location 6


Crash

PDO

Year 2011

Year 2010 613232 Sideswipe

(Same direc.)

PDO

Year 2009

Year 2008 8020230 RecfrPapSys Incapacitating

Injury

DBB Location 7


To Rear)

PDO


Year 2011

55


Year 2010 500656 RecfrPapSys Incapacitating

Injury




Year 2008 8062539 RecfrPapSys PDO

DBB Location 8


Crash

PDO

2645488 Sideswipe

(Same direc.)

PDO


Crash

PDO


Injury



To Rear)

Possible

Injury


Crash

PDO



Documents

Md Mozahidul Islam_MS Thesis