DEPARTMENT OF CIVIL & ENVIRONMENTAL ENGINEERING  
THE UNIVERSITY OF THE WEST INDIES 
 
ST. AUGUSTINE CAMPUS, TRINIDAD 
THE WEST INDIES. 
 
RESEARCH PROJECT FINAL REPORT 
ON 
EVALUATION OF THE NEW PEDESTRIAN CROSSING IN TRINIDAD 
By 
 
Dr. MADANIYO MUTABAZI 
 
 
October 2008
 i
TABLE OF CONTENTS 
  Page 
Table of Contents  i 
List of Tables  iii 
List of Figures  v 
List of Abbreviations, Acronyms, and Initialisms vi 
Acknowledgement vii 
Disclaimer vii 
Executive Summary viii 
CHAPTER ONE:  INTRODUCTION 1 
1.1 Background 1 
1.2 Problem Statement 1 
1.3 Study Objectives 2 
1.4 Study Methodology 3 
1.5 Scope of the Study 3 
1.6 Report Organisation  3 
CHAPTER TWO:  LITERATURE REVIEW 4 
2.1 Pedestrian Safety in Trinidad 4 
 2.1.1 Crash Data 4 
 2.1.2 Young Pedestrians 5 
 2.1.3 Pedestrian Infrastructure 7 
 2.1.4 Road Users’ Attitudes and Behaviours 7 
 2.1.5 Government Role 8 
2.2 The New Zebra Crossing 10 
 2.2.1 Historical Background 10 
 2.2.2 Zebra Markings 13 
 2.2.3 Advance Sign  17 
 2.2.4 Flashing Lights on Black & White Posts 17 
 2.2.5 Zigzag Lines  21 
 2.2.6 Safety Performance 23 
2.3 State–of-the Practice on Mid-block Pedestrian Crossings 24 
 2.3.1 Improved Mid-block Zebra Crossings 24 
 2.3.2 Traffic Controlled Mid-block Pedestrian Crossings 27 
CHAPTER THREE: METHODOLOGY 29 
3.1 Observed Drivers’ Yielding Behaviours 29 
 3.1.1 Experimental Design 29 
 3.1.2 Data Collection 30 
 3.1.3 Response Variable 31 
3.2 Drivers’ Opinion 32 
 3.2.1 The Questionnaire 32 
 ii
   Page 
 3.2.2 Demographic Factors 33 
 3.2.3 Drivers’ Understanding of Crossing Elements 33 
 3.2.4 Self-reported Yielding Behaviour  34 
 3.2.5 Estimate of Pedestrian Crossing Problem 34 
 3.2.6 Effectiveness of the New Zebra Crossing 34 
3.3 Field Visits and Pedestrian Survey 35 
CHAPTER FOUR: ANALYSIS AND DISCUSSION 36 
4.1 Observed Drivers’ Yielding Behaviour 36 
 4.1.1 Effect of Flashing Lights Operation 36 
 4.1.2 Effect of Pedestrian’s Walking Direction 38 
 4.1.3 Effect of Vehicle Size on Drivers’ Yielding Behaviour 40 
 4.1.4 Effect of Vehicle Use on Drivers’ Yielding Behaviour 41 
 4.1.5 Modelling of Driver’s Yielding Behaviour 43 
 4.1.6 Comparison with Other Studies 46 
4.2 Drivers’ Opinion 48 
 4.2.1 The Meaning of Zebra Markings 48 
 4.2.2 The Meaning of Flashing Lights 49 
 4.2.3 The Meaning of Zigzag Lines 51 
 4.2.4 Awareness of Traffic Laws at Zebra Crossing 53 
 4.2.5 Drivers’ Yielding Behaviour 54 
 4.2.6 Magnitude of Pedestrian Safety Problem 56 
 4.2.7 Effectiveness of New Zebra Crossing 56 
 4.2.8 Survey Reliability 57 
4.3 Field Visits and Pedestrian Survey 61 
 4.3.1 Pedestrian Views 61 
 4.3.2 The Crossing at Sevilla Private Primary School 63 
 4.3.3 Location of Priority-Pedestrian Crossing 64 
 4.3.4 Installation of Priority-Pedestrian Crossing 64 
 4.3.5 Motorists’ Behaviours 67 
CHAPTER FIVE: CONCLUSIONS & RECOMMENDATIONS 69 
5.1 Conclusions 69 
5.2 Recommendations 70 
REFERENCES 73 
APPENDIX 81 
 
 iii
LIST OF TABLES 
  Page 
Table 4.1 Summary of drivers’ observed yielding behaviour data 37 
Table 4.2 Effect of flashing lights on drivers’ yielding behaviour 37 
Table 4.3 Effect of pedestrian’s walking direction on drivers’ 
yielding behaviour 
39 
Table 4.4 Effect of flashing lights and pedestrian’s walking 
direction on drivers’ yielding behaviour 
40 
Table 4.5 Effect of vehicle size on drivers’ yielding behaviour 40 
Table 4.6 Effect of flashing lights within vehicle size on drivers’ 
yielding behaviour 
41 
Table 4.7 Effect of vehicle use on drivers’ yielding behaviour 42 
Table 4.8 Effect of flashing lights by vehicle use on drivers’ 
yielding behaviour 
42 
Table 4.9 Comparison with other studies on drivers’ yield rate 47 
Table 4.10 Descriptive statistics for survey respondents 48 
Table 4.11 Correlation between respondents’ demographic factors 48 
Table 4.12 Understanding of priority-pedestrian zebra crossing 
elements 
50 
Table 4.13 SPSS output on respondent attributes’ significance in 
predicting their understanding of the meaning of zebra 
markings. 
 
50 
Table 4.14 SPSS output on respondent attributes’ significance in 
predicting their understanding of the meaning of flashing 
lights 
51 
Table 4.15 SPSS output on respondent attributes’ significance in 
predicting their understanding of the meaning of zigzag 
lines 
51 
Table 4.16 Driver’s requirement for a pedestrian waiting to cross 54 
Table 4.17 SPSS output on respondent attributes’ significance in 
predicting their requirement for a pedestrian waiting to 
cross 
55 
Table 4.18 Reported frequency on drivers’ yielding for pedestrians at 
zebra crossings 
55 
   
 iv
  Page 
Table 4.19 Reasons why drivers not yielding for pedestrians at zebra 
crossings  
56 
Table 4.20 SPSS output on respondent attributes’ significance in 
predicting their support for the new zebra crossing 
57 
Table 4.21 Comparison between self-reported and observed drivers’ 
yielding behaviour 
59 
Table 4.22 Self-reporting and reporting others yielding behavior  59 
Table 4.23 Respondent’s consistency between “meaning of zebra 
markings” & “driver’s requirement at zebra crossing” 
60 
Table 4.24 Respondent’s consistency between “meaning of zebra 
markings” & “reason for not yielding to pedestrian at 
zebra crossing” 
61 
Table 4.25 Respondent’s consistency between “driver’s requirement 
at zebra crossing” & “reason for not yielding to pedestrian 
at zebra crossing” 
61 
 
 v
LIST OF FIGURES 
  Page 
Figure 2.1 New Zebra Crossing 11 
Figure 2.2 Flashing lights on black & white posts 12 
Figure 2.3 Advance sign at new zebra crossing 12 
Figure 2.4 Crossing obscured by adverse horizontal alignment (sharp 
horizontal curve) 
15 
Figure 2.5 Crossing obscured by adverse vertical alignment (sharp 
vertical curve) 
15 
Figure 2.6 Zebra markings bar length 16 
Figure 2.7 Crossing before re-marking 19 
Figure 2.8 Pedestrian trapped between opposite traffic on the 
crossing 
23 
Figure 3.1 Definition of pedestrian’s region 31 
Figure 4.1 Effect of flashing lights on drivers’ yielding behaviour 38 
Figure 4.2  Extract of SPSS output modelling drivers’ yielding 
behaviour 
45 
Figure 4.3 Zigzag markings used for advance warning 53 
Figure 4.4 Priority crossing without pavement markings 65 
Figure 4.5 Priority crossing without flashing lights 65 
Figure 4.6 Non-stripped post(s) 66 
Figure 4.7 Straight line painted over zigzag line 66 
Figure 4.8 Parking within the crossing 67 
Figure 4.9 Stopping within the crossing 67 
Figure 4.10 Overtaking within the crossing 68 
 
 vi
 
ABBREVIATIONS, ACRONYMS, AND INITIALISMS 
 
EMR—Eastern Main Road 
FHWA—Federal Highway Administration 
LED – Light Emitting Diode 
MoW&T—Ministry of Works & Transport 
MTC-- Metropolitan Transportation Commission  
MUTCD—Manual on Uniform Traffic Control Devices 
NMT—Non Motorized Transport 
NZ – New Zealand 
PBR—Priority Bus Route 
RC—Roman Catholic 
SPSS -- Statistical Package for Social Sciences  
SUV—Sport Utility Vehicle 
TCD—Traffic Control Device 
T&T—Trinidad & Tobago 
UK—United Kingdom 
 
 vii
ACKNOWLEDGEMENT 
Support for this project was provided by the Department of Civil & 
Environmental Engineering at the University of the West Indies, St. Augustine 
Campus. The authors would also like to acknowledge Mr. Howard Johnson, 
Deputy Principal Secretary Ministry of Works & Transport, Mr. Dexter WongWai 
& Ms Nadia Persard at the Traffic Branch of the Ministry of Works & Transport, 
Trinidad & Tobago for providing of important information, support, and 
comments on the draft report. A special thanks also to former undergraduate 
students Krystal Nanan and Dindial Rondell for their diligent efforts. 
 
DISCLAIMER 
The contents of this report reflect the views of the authors, who are responsible 
for the facts and accuracy of the data presented herein. The contents do not 
necessarily reflect the official views or policies of the Ministry of Works & 
Transport of the government of Trinidad & Tobago. This report does not 
constitute a standard, specification, or regulation.
 viii
Executive Summary 
 
Walking mode is widely used in everyday activities. Recently it has received 
more attention from transportation professionals because of its advantages over 
motorized transport modes especially for short trips. However, pedestrians are 
the most vulnerable road users for various reasons, more so in developing 
countries including Trinidad & Tobago. Most vehicle-pedestrian crashes in 
Trinidad are reported to involve a pedestrian crossing the roadway.  
 
A widely used conventional zebra crossing consisting of transverse white 
pavement markings across the roadway suffers low drivers’ yielding rates to 
pedestrians. Engineers have been searching for innovative engineering 
treatments to supplement zebra markings to increase driver’s yielding 
behaviour. In this endeavour the government of Trinidad & Tobago has 
introduced an improved zebra crossing in 2004, which will eventually replace 
conventional zebra crossing.  
 
Flashing lights on painted posts, zigzag lines on the pavement, advance sign, and 
pedestrian-friendly guidelines when using the crossing are the major 
improvements on this crossing compared with conventional zebra crossing. The 
new crossing has a unique image; visible from sufficient distance under all 
roadway, traffic, and weather conditions. These features are vital in increasing 
yielding behaviour from an average driver. 
 
The study involved evaluation of the new crossing, specifically the effect of the 
crossing on motorists; and public opinion about the crossing. The study did 
evaluate the effect of the crossing on pedestrian crossing behaviour. It was done 
through a survey administered to drivers, interviewing pedestrians, field 
observation of drivers’ behaviour at crossings during daylight hours.   
 ix
 
Flashing light operation was associated with increased drivers’ yielding 
behaviour observed in the field. However, self-reported drivers’ yielding rates 
were much higher than observed rates—self-reporting bias. Drier’s risk 
perception of potential collision with a pedestrian prompted driver’s yielding 
than activation of flashing lights. Pedestrians were not satisfied with drivers’ 
yielding behaviour probably because of initial higher expectations. Drivers 
support the new crossing to increase pedestrian safety. Most drivers do not 
understand the meaning and hence the purpose of zigzag lines at the new 
crossing, leading to frequent disrespect of zigzag lines prohibitions (parking, 
overtaking, and stopping) at the crossing. 
 
Engineering features of the priority crossing offers a viable solution for 
deficiencies of conventional zebra crossing, specifically of all-weather visibility. It 
promises in improving pedestrian safety in Trinidad & Tobago, and is likely to 
benefit developing countries where driving culture marginalizes pedestrians and 
road maintenance practices are not adequate. However, based on experience 
from other countries there exist opportunities of improving the crossing. 
Engineering, Education, Enforcement, and Research elements are recommended 
for further improvement of the crossing. 
 1
CHAPTER ONE: INTRODUCTION 
1.1 Background 
More emphasis is being placed on Non-Motorised Transport (NMT) modes 
(walking and cycling) for short trips. NMT modes are more healthy and 
environmental friendly than motorised modes. NMT support safer streets, 
cleaner environment, and increased activity on the part of the traveller, which 
may help to fight some common health problems such as diabetes, etc. Walking 
mode is used by almost every trip maker, be walking to parking lots, to public 
transport stands/stations, or even as the only mode for the entire trip.  
 
Besides advantages offered by the walking mode, still it is considered a risky 
undertaking. Road safety statistics indicate pedestrians as the most vulnerable 
road user group. The problem is of higher magnitude in developing countries, 
including Trinidad & Tobago (T&T).   High frequency of pedestrian-related 
crashes is a result of higher pedestrians’ exposure to vehicular traffic, mostly 
when crossing roadways; and inability of pedestrians’ bodies to absorb impact 
energy during crashes. In recognition to the pedestrian safety problem in T & T, 
the Ministry of Works & Transport (MoW&T) in 2004 initiated installation of 
new pedestrian zebra crossing type which will eventually replace conventional 
zebra crossing.  
 
Zebra crossings are places where pedestrians can cross the road legally on the 
same level/grade as vehicular traffic. They are marked by black & white strips 
across the pavement with or without supplemental engineering design elements 
or/and traffic control devices. They could be either at intersections or at mid-
block sections.  
 
1.2 Problem Statement 
For the purpose of this study we define a conventional zebra crossing to consist 
of only alternate transverse white pavement markings across the roadway 
 2
without any engineering design modification or supplemental traffic control 
device(s) (TCD). Significant proportion of drivers who don’t yield the right-of-
way to pedestrians at conventional zebra crossings has been found to be a major 
setback for these crossings. As low as five percent drivers’ yield rate have been 
observed at zebra crossings at mid-block location (Várhelyi, 1998). In another 
study by the same author, 73% of drivers either kept or increased their speed 
when approaching a zebra crossing in the presence of a pedestrian (Várhelyi, 
1996). Both parties involved in a pedestrian related crash possess opposing views 
and attitudes toward pedestrian safety. While motorists blame pedestrian for 
risky behaviours, pedestrians blame motorists for marginalizing pedestrians. 
Redmon (2003) asserts that these attitudes change depending on whether one is a 
driver or a pedestrian.  
 
Unsatisfactory yielding behaviour by drivers to pedestrians at conventional 
zebra crossings has prompted most road agencies to search for innovative 
engineering treatments to supplement zebra markings (Dun, 1989; Public Works, 
2003). Some of these treatments have been successful than others in increasing 
drivers’ yielding rates (Public Works, 2003). Effectiveness of road safety projects 
and public opinion about these projects plays a significant role in sustaining such 
project for long term. We couldn’t find any literature addressing the application 
and effectiveness of the new zebra crossing adopted in T&T. In light of the 
above, there is a need for evaluation of this new crossing. 
 
1.3 Study Objectives 
The main objective of the study was to evaluate the new pedestrian crossing, 
recently introduced in T&T, specifically its: 
i) Effectiveness on increasing drivers’ yielding behaviour to pedestrians, and 
ii) Public opinion on the new crossing. 
 
 
 3
1.4 Study Methodology 
The study involved literature review; triangulation method about drivers’ 
yielding behaviour (public opinion, self-reported & observed), and public 
opinion on the effectiveness of the new crossing. Also field visits to the new 
crossings was made to get familiar with the crossing application and use. 
 
1.5 Scope of the Study 
This study is limited to evaluation of the new zebra crossing in Trinidad. The 
crossings are located at mid-block sections. The study didn’t include effect of 
crossing on pedestrians’ behaviour. Driver’s behaviour at the crossing was 
observed only during daylight hours. The terms “crosswalk” and “crossing” are 
used interchangeably in this report.  
 
1.6 Report Organization   
The report is made up of five chapters. Chapter one deals with introduction such 
as background information, problem statement that lead to the formulation of 
the project, brief discussion on study methodology, and scope of the study. 
Chapter two scoops the literature to solicit information relevant to mid-block 
zebra pedestrian crossings. Chapter three presents study methodology and data 
collection, while chapter four presents data analysis and discussion of the results. 
Chapter five caps the report with conclusions drawn from the study and 
suggestions to improve pedestrian safety at these crossings in T&T. 
    
  4 
CHAPTER TWO: LITERATURE REVIEW 
Available literature was searched on the subject of pedestrian safety at mid-block 
zebra crossings. 
2.1 Pedestrian Safety in Trinidad 
Similar to most developing countries, pedestrians and cyclists in T&T are the 
most vulnerable road users. It is reported that between 1965 and 2000 
pedestrians have constituted between 42 and 50 percent of all road fatalities and 
33-41 percent of all road injuries in T&T. While pedestrians have traditionally 
accounted for the majority of injuries and road fatalities since 1960, decline of 
cyclists injured and killed have been observed. (St. Bernard and Mathews, 2003). 
 
“Pedestrian crossing the roadway” is the frequently reported scenario in crashes 
involving pedestrians in Trinidad. Pedestrian infrastructures are inadequate, 
under-utilised and misused, while most pedestrians are considered high-risk 
takers. Local driving culture which is understood to marginalize pedestrians has 
a major influence on pedestrian safety. Pedestrian safety in Trinidad is hereby 
summarised by addressing crash data, young pedestrians, pedestrian 
infrastructure, road users’ attitude and behaviour, and the government’s role on 
pedestrian safety. 
 
2.1.1 Crash Data 
Official data on crashes involving pedestrians are collected and maintained by 
the Traffic Branch of the Police Department in the Ministry of National Security. 
Most roadway attributes at the crash site, specifically, the geographical locations 
are too general for detailed research purposes. Although in most reported 
pedestrian crashes, “pedestrian crossing the roadway” is the frequent reported 
scenario, some members of the public may not always agree to this reason. 
Limitations of crash data make it difficult to ascertain the pre-crash conditions 
and crash locations in relation to marked crosswalks. 
    
  5 
 2.1.2 Young Pedestrians 
Although adults make up the largest percentage of pedestrian crashes, young 
children’s safety, as pedestrians, is of particular concern in view of their 
vulnerability in traffic situations and the special value societies places on 
children. For the years 1989—2002 in Finland it was found that at sites with no 
marked pedestrian crossings, children were more often injured, compared to the 
other age groups. Similar trend was observed at mid-block, marked crossings on 
multilane sections due to overtaking vehicles (Leden et al., 2006). Young 
pedestrian vulnerability stems from a number of factors including their smaller 
stature, small and overestimated cognitive skills, their unpredictable actions, lack 
of experience as road users, and their inability to drive making them heavily 
dependent on NMT modes including walking.  
 
Their unpredictable actions are the result of the fact that they may be old enough 
to be mobile but too small to be easily visible from the driver’s position. They 
have difficulty of seeing and evaluating the entire traffic situation correctly 
because of their short height. They have difficulty in correctly perceiving the 
direction, sound and speed of vehicles. They are also easily distracted. Redmon 
(2003) reported that one of drivers’ concerns is the possibility of children running 
in front of them. A study of children crossing road at intersection found that less 
than half of the children looked in the direction of oncoming traffic before 
crossing, while a full visual search (looking in all directions before and while 
crossing) was carried out by fewer than 5 percent (MacGregor et al., 1999). 
Connelly et al., (1998) found that children between ages of 5 to 12 years use the 
same distance gap threshold in road crossing decisions irrespective of speed of 
the oncoming vehicle. 
 
Adults tend to overestimate children’s pedestrian skills. The extent of 
overestimation of children’s road use skills by parents increased with decreasing 
    
  6 
child’s age. However, estimation of child’s vocabulary skills by the same parents 
was not influenced by child’s age (Paediatrics for Parents, 1991). Children’s 
ability overestimation was also observed by Harré (2003) in the study of drivers’ 
speeds in the presence of child, in which New Zealand drivers were found to 
make inadequate speed adjustments in the presence of children. Similar results 
were obtained in United Kingdom (UK) where the presence of children by the 
roadside had no effect on drivers’ speeds or lateral positions of their vehicles 
(Thompson et al., 1985). 
 
Many children believe that the safest way to cross the street is to run, that it is 
safe to cross against the red light, and that adults will always be kind to them by 
stopping instantly if they are in danger (MacGregor et al., 1999). The same study 
found that a number of gaps exist between the instructions parents give, and 
their expectations of children’s capabilities, and what children actually do when 
crossing roadways. 
 
Studies on school transport conducted to public schools in St. George East 
County in north-central Trinidad revealed an increased percentage of school 
children who walk to school between 1980 and 2003 (Mutabazi and Akberali, 
2008; Underwood and Rasul, 1984). Proportion of younger children (from 
primary schools) walking to school is higher than that of corresponding older 
children (from secondary schools) (Mutabazi and Akberali, 2008). This implies 
that more young school children, as pedestrian, are exposed to vehicular traffic 
than older school children, despites the fact that older children are more likely to 
make safe crossing decision than younger children (Connelly et al, 1998). 
 
Crossing roadways while going to, or coming from school is not limited to only 
children who walk to school as the main mode of transport. Those who use 
public transport are also required to walk and/or cross roads to get into school 
compounds, the extent of such activities depend on the relative location of the 
    
  7 
school compound and public transport vehicles stops. Students being picked and 
dropped by parents/guardians using private transport, sometimes may also be 
required to cross roadways to get into school compounds due to inadequate 
parking facilities at most schools in the country. Recognizing the need for school 
children safety, MoW&T prioritized school areas when installing new zebra 
crossings. 
 
2.1.3 Pedestrian Infrastructures 
Pedestrian infrastructures are inadequate, under-utilised and misused. 
Pedestrian grade-separated crossings (pedestrian bridges) are few because are 
expensive to construct. Use of pedestrian bridges without escalators in 
developing countries such as T&T is not always encouraging (Räsänen et al., 
2007). As such most pedestrians are seen crossing vehicular traffic streams on 
busy roadways. Sidewalks on most parts of urban and suburban streets are 
either non-existent or are insufficiently planned and maintained. It is common to 
find a sidewalk crossing the roadway and yet the crossing is not painted 
(unmarked crossing) or the sidewalk being terminated abruptly along its route.  
 
2.1.4 Road Users’ Attitudes and Behaviours 
Pedestrians are normally considered high-risk takers. A study of a level crossing 
over the railway track used by both vehicles and pedestrians in Finland found 
that pedestrians and cyclists had higher frequency of violation than motor 
vehicles (Knutton, 2004). Pedestrians’ unsafe behaviour in Trinidad includes 
crossing roadways at non-authorized locations. Authorized crossings are 
considered to include pedestrian over-/under-passes, and marked crossings. 
Other unsafe behaviours include violating traffic signals, walking on roads (jay 
walking), etc. This has prompted some members of the public and government 
officials to suggest that pedestrians should be accountable for their behaviours 
when using roadways. 
 
    
  8 
Local driving culture has a major influence on pedestrian safety. Drivers’ 
behaviours that affect pedestrian safety in Trinidad include non-yielding to 
pedestrians crossing roads; parking vehicles on footpaths and sidewalks; 
encroaching pedestrian areas without due consideration to pedestrian when 
drivers are trying to avoid rough pavement surfaces such as potholes; and hit-
and-run attitude, especially when pedestrians are involved in crashes. In some 
cases running away drivers involved in such crashes may claim that they 
thought they hit objects other than pedestrians. Hit-and-run behaviour can 
present difficulties in processing compensation from insurance companies to 
crash victims and their families. 
 
Misuse of pedestrian facilities by sidewalk vendors, intersection petty traders, 
and commercial enterprises can degrade pedestrian safety. While sidewalk 
vendors conduct their activities on pedestrian sidewalks, business community 
considers sidewalks to be part of their parking lots. These behaviours in 
conjunction with physical poor condition of sidewalks’ surfaces result into 
pedestrians walking in middle of roads creating safety problems. Intersection 
petty traders apart from exposing themselves to traffic crashes they are source of 
drivers’ distraction resulting into crashes. 
 
2.1.5 Government Role 
To some extent road users’ unsafe behaviours reflects the unbalance between 
what a transportation system offers (supply) and the demand placed upon it.  
For example, parking vehicles on sidewalks in part could reflect inadequate 
parking spaces and/or weakness in parking regulations & control. Mitigation of 
most road safety problems requires both an increased capacity to road 
infrastructures and regulations on how road users should behave on roadway 
environment. In Trinidad regulation outpace capacity improvement probably 
because the former requires minimum resources to set them up. However, some 
    
  9 
of regulations that are in place may not be effective because of some inherited 
weaknesses. 
 
Enforcement of traffic regulations related to” pedestrian crossings” are either 
missing or inadequate. Statuary regulations supporting right-of-way for 
pedestrian at zebra crossing is too general to be enforced in an effective manner. 
The law requires motorists to yield to pedestrians, not specifically stopping. The 
only proof that the motorist did not yield would be when the pedestrian is hit. It 
is only recently when introducing the new zebra crossings that drivers have been 
directed to stop for pedestrians. The fact that the directive is considered a mere 
guideline and is not part of the law, its enforcement is questionable. Some 
enforcement agencies use stopping condition or speed limit, as a criterion for 
drivers’ violation at a crosswalk. These measurable criteria make enforcement 
feasible and easier. Enforcement of regulations pertaining zebra crossing in 
Trinidad cannot be effected efficiently until speeding and stopping criteria are 
part of the pedestrian crossing laws. 
 
Apart from providing roadway infrastructure and setting up associated 
regulations, the government have the role of setting up supportive programs. 
One such program is school safety for children who are more at risk of road 
crashes, as discussed in section 2.1.2 above. School speed zoning, road safety 
education for the young road users, traffic calming in school zones, safe 
pedestrian routes to school, and crossing guards would have equipped school 
children with appropriate tools and means for walking and cycling safely to/and 
from schools.   
 
Lack of public policy may jeopardize pedestrian safety. A motorist complaining 
against having a crossing guard to stop traffic to help pedestrian cross the 
highway asserted that the Minister was against such exercise because it blocks 
traffic on that highway (Express
a
, 2004). Yet in another instance, the local 
    
  10 
newspaper carried out a story of a traffic light signal installed by the government 
outside Dede High School to assist pedestrian cross the busy road. However, the 
traffic light signal could not put into use for two years because no one was ready 
to pay initial cost for electricity (Express
b
, 2005). These issues may in part reflect 
a policy problem. 
 
2.2 The New Zebra Crossing 
This section describes the new zebra crossing based on information obtained 
from literature survey and authors’ experience. The new zebra crossing consists 
of a conventional zebra crossing supplemented with zigzag lines, yellow flashing 
lights sometimes called flashing beacons or simply flashers, and an improved 
advance warning sign. 
 
2.2.1 Historical Background 
Literature indicates that the new zebra crossing in T&T was officially 
documented in 1993 (Government of Trinidad & Tobago
a
, 1993). Figure 2.1 
shows the early documented crossing on which pedestrians are prohibited from 
crossing the carriageway except within the limits of the crossing’s give-way lines. 
It is made up of: 1) traditional black and white stripes (zebra stripes) parallel to 
the curb—i.e., the conventional zebra crossing, 2) longitudinal zigzag lines 
parallel to the roadway on both approaches, and 3) flashing yellow globes 
mounted on top of a white-and-black stripped poles placed on the side of the 
roadway at the crosswalk locations. The globes flash continuously at a rate of 
one-per-second. In UK this globe is known as orange belisha beacon named after 
Leslie Hore-Belisha, the Minister of Transport who introduced them in 1934. In 
New Zealand yellow belisha beacons flashes only at night (Dunn, 1989), and they 
could be substituted by fluorescent beacon discs (Land Transport NZ, 2006). The 
crossing has some similarities to the crossing used in the United Kingdom (UK), 
New Zealand (NZ), and in the former British crown colonies of Singapore, South 
    
  11 
Africa, and Hong Kong. There is no evidence that this crossing ever got 
implemented until 2004. 
 
The crossing was amended in 2001 by:- 1) replacement of a yellow globe with 
two circular amber (yellow) wigwag signals facing opposite directions of traffic 
(Figure 2.2), and 2) explicitly defining the right-of-way for a pedestrian at the 
crossing (Government of Trinidad & Tobago
b
, 2001). It is referred to as priority-
pedestrian crossing due to its explicit definition of the right-of-way to a 
pedestrian. A conventional advance warning sign was amended with additional 
of words “ZEBRA CROSSING” (Figure 2.3) in 2004 when the first series of 
pedestrian-priority zebra crossings were launched (Guardian, 2004).  
 
Figure 2.1: New zebra crossing (Government of Trinidad & Tobago, 2001
b
) 
    
  12 
  
 
 
The Ministry of Works and Transport gave guidelines for the use of the crossing, 
summarized as follows: - (Guardian, 2004) 
1) The pedestrian has the right-of-way on pedestrian-priority zebra crossing, 
and therefore the driver must stop to permit the pedestrian to cross the 
roadway; 
2) A maximum speed of 25 Km/h when approaching the pedestrian-priority 
zebra crossing; 
3) Although pedestrians have right-of-way on “pedestrian-priority zebra 
crossings” still they should exercise care and caution when using them. 
 
Some members of the public understood these guidelines to be the first to accord 
a pedestrian the right-of-way when crossing the roadway at zebra crossings in 
the country (Defour, 2004). Prior to these guidelines pedestrian were accorded 
right –of –way only at pedestrian crossings controlled by light signal or by the 
police, according to article 23 of the Highway Code (Ministry of Works, 1972). 
Right-of-way guideline has an element of regulation and is intended to ensure 
there is no doubt where and when a vehicle should stop at a pedestrian-priority 
Figure 2.2: Flashing lights on black & white posts 
(Government of Trinidad & Tobago, 2001
b
) 
Figure 2.3: Advance sign at 
new zebra crossings 
    
  13 
zebra crossing. Speed limit guideline puts the driver in an environment that 
he/she can stop easily. However, we found that these two measures are only 
documented in newspapers -- this leads to a question on their statuary power. 
The third measure is intended to protect pedestrians from being over-confident 
at these crossings, a phenomenon that may lead to unsafe crossing behaviour on 
the part of pedestrians.  
 
2.2.2 Zebra Markings 
Transverse pavement (zebra) markings are intended to define the area where a 
pedestrian can cross the road legally, in a way; pedestrians are channelled and 
guided into the proper path. They help the driver to see the crossing well in 
advance. The combination of these two features is expected to improve the safety 
of pedestrians using the crossing. However, study results are mixed and 
inconclusive pertaining to the safety effect of marking crossings at uncontrolled 
locations. Campbell et al. (2004) after reviewing pedestrian safety research 
concluded that marked crosswalks at uncontrolled intersections on two-lane 
roads are not necessarily safer than unmarked ones unless they are accompanied 
with some additional engineering improvements. Koepsell et al. (2002) 
concluded that crosswalk markings appear associated with increased risk of 
pedestrian-motor vehicle collision to older pedestrians at sites where traffic is 
controlled. Zegeer et al. (2004) asserts that these studies were inconclusive 
because of inherited limitations in methodological approach, sample size 
limitation, and limitation on accounting contributing factors.  
 
Wide application of zebra markings at pedestrian crossings worldwide indicates 
the extent road agencies’ belief on their effectiveness. Ekman and Hyden (1999) 
argue that despite the well-known fact that most of the accidents where 
pedestrians have been hit by a car in urban areas occur either at a zebra crossing 
    
  14 
or at a signalized intersection; this has not changed the opinion by many people 
about the zebra crossing being a safer place to cross the roadway.  
 
Zebra markings are only useful if they are visible at a distance far enough to 
allow the driver to take safer action. Some agencies use zebra markings visibility, 
among the criteria for deciding the zebra crossing location (Vermont Agency of 
Transportation, 2004). Zebra markings visibility depends on roadway geometric 
design; marking material reflectorization property, weather, and road 
maintenance level & practice. 
 
Roadway Geometric Design 
Three geometric design elements that affect the visibility of zebra markings are 
horizontal alignment, vertical alignment, and overall width of zebra markings 
(i.e. the length of crossing in driver’s line of sight). Unfavourable horizontal 
and/or vertical alignment may result into non-visibility of the crossing from 
driver’s position at a sufficient distance. Figure 2.4 & 2.5 demonstrate sight 
distance restriction due to adverse horizontal and vertical alignment 
respectively. Location decisions of pedestrian crossings to solve roadway 
alignment problem may at times conflict with other location criteria such as 
spatial relationship between pedestrian crossing and pedestrian traffic 
generators. 
 
Even with perfect horizontal and vertical alignments, yet the driver may be 
unable to see the markings at sufficient distance due to smaller width of zebra 
markings. The required width depends on driver’s eye height, vehicle 
approaching speed, and driver’s minimum angle of resolution. New Zealand 
using a model shown in Figure 2.6 with driver’s eye height (h) of one meter, 
driver’s minimum angle of resolution (?) of two minutes, and appropriate 
stopping sight distances, for each design speed, they developed desirable zebra 
bar lengths (l). For example at speeds of 60 & 70 Km/h desirable length l should 
    
  15 
be 5.6 & 9.10 metres respectively (Dunn, 1989). Values recommended by Litman 
et al. (2002) are 3.0 & 4.0 metres respectively.  Model used by Litman et al. could 
not be found in the literature. 
 
 
Figure 2.5: Crossing obscured by adverse vertical 
alignment (sharp vertical curve) 
Figure 2.4: Crossing obscured by adverse horizontal alignment (sharp 
horizontal curve) 
    
  16 
 
Wider markings may reduce pavement skid resistance, making it slippery when 
wet for motorists. Minimum width of markings by both Dunn and Litman 
requires wider painted area at higher speeds when skid more resistance is 
needed. Recognizing this problem, New Zealand recommends a cut-off value of 
five metres, and encourages the use of mono-plastic/spray plastics rather than 
paint (Dunn, 1989). T&T guidelines on the minimum overall width of zebra 
markings are not known if they exist.  
 
Marking Material Retroreflectorization 
Zebra markings usually fade away with time, traffic, environmental condition, 
and other factors (Smadi et al., 2008) making them difficult to see at sufficient 
distance. The quality and type of the road marking material have a significant 
influence on marking’s durability and retroreflectivity. Lee et al. (1999) asserted 
that solvent borne paint, which is mostly used by MoW&T has a lowest 
durability period of less than one year compared with other road marking 
materials such as thermosets, thermoplastics and tapes which can last up to 
three, five, and eight years respectively. When to re-mark zebra markings is an 
issue of road agency maintenance level and practice. Most agencies schedules re-
marking of pavement markings based on experience.  
 
D = Stopping sight distance 
l =Bar length 
h = Driver’s eye height 
? 
? = Driver’s minimum resolution 
 h 
 l D  
Figure 2.6: Zebra marking bar length (Dunn, 1989) 
    
  17 
Inclement weather such as fog, haze, heavy rain, etc. limits how far a driver can 
see. At night time on unlit crossing, the source of light is the vehicle’s beam light. 
By then visibility of zebra markings depends on the amount of beam light 
reflected back by the marking’s, which in turn depends on strength of the 
original beam light and retroreflective characteristics of the pavement marking 
material. Amount of light reflected back to the driver depends on the angle made 
up by driver’s line of sight and vehicle’s beam light -- observation angle (Smith, 
2008). Drivers’ eyes are further from the headlamps in larger vehicles, resulting 
in a greater observation angle and hence lesser amount of reflected light back to 
the driver. While conscious design and precise fabrication of the prismatic 
structures used in modern sheeting can solve this problem for signs, it is not the 
same for road paints.  
 
2.2.3 Advance Sign 
Some local agencies would normally supplement zebra markings with signing(s) 
and/or pavement markings in the advance zone to enhance crossing visibility. 
Litman et al. (2002) recommends that zebra markings should be supplemented 
by signs as the means to identify the place of the zebra crossing. The advance 
sign for the pedestrian-priority zebra crossing (Figure 2.3) intends to alert the 
driver on the approaching of zebra crossing. Additional words on the sign serve 
to differentiate pedestrian priority zebra crossing from a conventional zebra 
crossing. However, a sign, posted at height above the region of vehicle beam 
lighting, may become difficult to be read at a sufficient distance during 
nighttimes (Rys et al. 2001).  In busy urban roads, meaning of advance sign or 
marking could get lost in the clutter and hence of little importance. 
 
2.2.4 Flashing Lights on Black & White Posts 
Yellow flashing lights also known as yellow flashers or beacons are mounted on 
posts painted with alternate black & white reflective bands. Flashing lights 
    
  18 
supplement pavement markings and advance sign in identifying the position of 
the crossing. Yellow flashing lights have also been used:- 
• In school zones;  
• In median island noses to provide motorist guidance of obstacles;  
• On construction barricades where there are lane reductions, detours, or other 
unexpected changes in traffic conditions;  
• At sharp horizontal curves to influence drivers’ speed;  
• At rail-highway grade crossings to indicate the presence of the train; and 
• At Kaola pedestrian crossing in South Australia.  
Twin yellow flashing lights at Kaola crossing generally operate during times 
when most schoolchildren are travelling to and from school; in the mornings and 
afternoons of school days only. 
 
Installation  above the ground and stronger visual cue, yellow flashing lights 
possess advantages over signs & pavement markings in attracting drivers’ 
attention making the crossing visible at sufficient distance in cases of:- 
• Awkward horizontal or/and vertical alignment; 
• Shorter zebra marking lengths; 
• Inclement weather & night time;  
• When pavement markings are obscured by traffic, or have faded away; 
• During the time after resurfacing and before re-marking is done (Figure 2.7); 
• Also where maintenance level or maintenance practise of traffic control 
devices is inadequate leaving signs and markings in a worn state most of the 
time. 
    
  19 
 
 
Flashing Light Features 
Effectiveness of flashing lights depends on its size, quantity, colour, flashing rate, 
type of light, height position, mode of operation, and configuration. The crossing 
uses four (two on each post at either side of the crossing) amber incandescent 
circular signals 300 mm in diameter, continuously flashing at the rate of 60 cycles 
per minute, positioned between 2.1 to 2.6 m from the ground. The two units on a 
post flash alternatively in a wigwag fashion similar to rail-highway grade 
crossing. Two flashing lights, at each side, both face the same on-coming traffic, 
ads an advantage to the crossing in case one of them is malfunctioning. 
 
The size of 300 mm in diameter used at this crossing was found to increase 
conspicuity compared to the 200 mm at a rail-highway grade crossing (Ruden 
and Coleman, 1979). In the same study, placement position of flashers at 2.7 m 
from the ground level was better than 5.1 m in terms of conspicuity. The upper 
limit height of 2.6 m used in Trinidad is close to 2.7 m tested by Ruden and 
Figure 2.7: Crossing before re-marking  
    
  20 
Coleman. However, the effect of placement height lower than 2.7 m could not be 
found in literature. Faster flash rate than that used at this crossing has been 
found to increase conspicuity while strobe was better than incandescent in 
Ruden and Coleman (1979) study. Faster flashing Light Emitting Diode (LED) 
beacons increased drivers’ yielding behaviour in Florida (Shurbutt et al., 2008). 
 
A study that compared several treatments to improve motorist yielding to 
pedestrians at un-signalized intersections indicated that red signal or red beacon 
devices produced higher yielding behaviour than in-roadway signs, yellow 
overhead flashing beacons, pedestrian crossing flags, and in-roadway warning 
lights (Turner et al., 2006). In the laboratory study of flashing lights at railway 
grade crossing, red flashing lights were found better in attracting subjects 
(Ruden and Coleman, 1979). The City of Los Angeles uses mid-block pedestrian 
signals that display a flashing red signal when activated (Fitzpatrick et al., 2006). 
 
It appears that pedestrian actuated flashers prompt drivers to yield for 
pedestrian more than continuous flashing mode used in Trinidad. Increased 
driver yielding to pedestrians at crosswalk with actuated flashers in 
Chattanooga, Tennessee, was attributed to flashers actuation (Van Winkle and 
Neal, 2000). Fitzpatrick et al. (2006) reported several studies that have shown that 
intermittent flashing beacons provide a more effective response from motorists 
than continuously flashing beacons. Sparks and Cynecki (1990) found that 
continuous flashers offer no benefit for intermittent pedestrian crossings in an 
urban environment and conceded that actuated warning flashers may be 
beneficial in a high-speed rural environment with unusual geometrics, high 
pedestrian crossings, and unfamiliar drivers. Some agencies prefer activated 
flashing over continuously flashing beacons, for the fear that the later may 
eventually lose its effectiveness because it becomes like a background to a driver 
who has driven through the section for a long time. 
    
  21 
Posts Stripped in Black & White 
It is a common practice in Trinidad to paint sign posts with black & white strips. 
In the study of low-cost traffic control systems for passive rail-highway grade 
crossings, Russell and Kent (1993) found that reflectorization of a crossbuck sign 
along with its post had a high visual impact on driver behaviour at rail-highway 
grade crossing. Similarly Yeh and Multer (2007) argue that the fully reflectorized 
posts allow drivers to see where the roadway meets the railroad tracks. 
Reflective black & white strips on the posts would reflect back the beam light 
from the vehicle at night, thereby enable the driver to locate the position of the 
crossing in case flashing lights are not working due to malfunctioning or power 
cut. This way the position of a crossing is easily identified during all weather 
conditions.  
 
2.2.5 Zigzag Lines 
Zigzag lines are marked on both lane edges and sometimes along the roadway 
centre line. Zigzag lines at a zebra crossing are also used at zebra crossings in 
UK, South Africa, New Zealand, and some agencies in Australia. The practice in 
Australia is to mark only along the centre line on both approaches.  
 
Purpose of zigzag lines in T&T, UK and South Africa is to restrict parking, 
overtaking, and stopping of vehicles within the crossing. In South Africa they 
also restrict changing lanes, and pedestrians prohibited from crossing within the 
zigzag zone (Ribbens, 1996). In Australia however, zigzag lines are used as an 
advance warning to the pedestrian crossing. Restricting parking, overtaking and 
stopping in the vicinity of a crossing can increase the visibility of a pedestrian 
entering the crossing. This can reduce the potential of an accidents resulting from 
hidden pedestrians by parked/stopped or overtaking cars (multiple threat).  
Advance yielding by shifting a stop line upstream the pedestrian crossing was 
one of the mitigation measures showed pedestrian safety improvement at 
    
  22 
pedestrian crossings (Van Houten et al., 2001; Van Houten et al., 2002) that. In 
Finland, children pedestrians were more vulnerable to crashes with overtaking 
vehicles at mid-block, multilane, marked crossings (Leden et al., 2006). In New 
Zealand, Canada and Europe parking or stopping too close to the crossing was 
found to be the most significant safety problem (Dunn, 1989). Another negative 
effect of parked/stopping vehicle at the crossing would be the obscure of 
flashing light(s) by big trucks, tall enough to obstruct the visibility of flashing 
light(s).A typical example which zigzag lines intend to avoid is shown in Figure 
2.8 where a pedestrian is crossing in front of stopped mini-bus may be difficult to 
see from a vehicle that would try to overtake the min-bus. 
 
In Trinidad zigzag lines extend to a maximum distance of about sixteen metres 
on either approach of the crossing. Standards allows as low as four metres in case 
the crossing is too close to the intersection. In South Africa they extend a distance 
of 30 metres on either approach (Ribbens, 1996). In Los Angeles a length of no-
stopping zone at pedestrian crosswalk is based on the safe stopping distance 
(Jones and Tomcheck, 2000). It appears the no-parking zone at priority 
pedestrian crossing in T&T, defined by the length of zigzag lines, is too short 
compared with other agencies. 
  
    
  23 
 
 
2.2.6 Safety Performance 
Road user’s expectation plays a key role in road safety. Drivers expect 
pedestrians at crosswalks that are located at intersections more than at 
crosswalks located at mid-block locations. Likewise, pedestrians expect drivers 
to yield to pedestrians on crossings at intersections more than those at mid-block 
locations. In a survey conducted by Ullman et al. (2004) concluded that 
unpredictability of a driver as to whether he/she will yield to a pedestrian at a 
pedestrian crossing was the main concern to pedestrians.  
 
Another factor that influences pedestrian safety at at-grade pedestrian crossings 
is the level of traffic control. For the purpose of this report, the term “traffic-
controlled” refers to pedestrian crossing where vehicular traffic is regulated by 
“STOP sign” or traffic signals. At such locations, driver’s yielding behaviour is 
improved by TCD’s command rather than drivers’ discretion. This phenomenon 
is likely to make the crossings at traffic-controlled locations safer than those at 
uncontrolled locations (Dunn, 1989). Chu et al. (2002) developed a model of how 
Figure 2.8: Pedestrian trapped between opposite traffic on the crossing 
    
  24 
pedestrians cross roads. From the model it was found that people are more likely 
to cross at a traffic-controlled intersection. Also, people are more likely to cross at 
any location with a marked crosswalk than at those without. The presence of a 
marked crosswalk is more influential at an intersection than at a mid-block 
location. Most of the priority zebra crossings in Trinidad are at uncontrolled mid-
block locations on two-lane roads. 
 
2.3 State-of-the Practice on Mid-block Pedestrian Crossings 
Mid-block pedestrian crossings are classified into three categories: 1) 
conventional zebra crossing, 2) improved zebra crossing, and 3) Traffic 
controlled pedestrian crossings. Conventional mid-block zebra crossings consists 
only zebra markings. Improved zebra crossing is a conventional zebra crossing 
supplemented by one or more passive traffic control devices or engineering 
design features. Traffic controlled pedestrian crossing normally do not contain 
zebra markings, instead they have traffic signals to control vehicular traffic as 
well as pedestrian signals; and they may contain also some improvement 
features used at improved zebra crossing. 
 
2.3.1 Improved Mid-block Zebra Crossings 
Improved zebra crossings consist of conventional zebra crossing supplemented 
by one or more engineering elements in the form of traffic control devices or 
roadway design elements. They are intended to:- 
• Improve visibility of crossing & pedestrian from drivers’ positions, 
• Remind drivers & pedestrians, 
• Reduce vehicle speeds, 
• Reduce vehicle-pedestrian conflicts & pedestrian exposure, 
• Provide pedestrian detection system. 
 
Crossing & Pedestrian Visibility  
Crossing visibility improvement features may include:-  
    
  25 
1) Passive or active signs placed either in the advance zone or at the crossing (at 
the sides or overhead) informing driver the location of the crossing. Some 
overhead signs are internally illuminated to increase conspicuity, while 
others may project incandescent light downward onto the crosswalk which 
helps motorists see pedestrians in the crossing at night. Nitzburg and 
Knoblauch (2001) concluded that the high-visibility crosswalk treatments had 
a positive effect on pedestrian and driver behaviour on the relatively narrow 
low-speed crossings. 
2) Advance signs and pavement markings in the form of zigzag lines—an 
Australia practice, diamond shape—New Zealand practice, or words on 
pavement, etc.  
3) Flashing lights or belisha beacons flashing continuously or activated by 
pedestrians. They use different flash rates, sequences, with/without strobe 
effect. Flashing lights sometimes placed in overhead position or in-pavement.  
4) Conspicuous reflectors such as fluorescent disc on a post in New Zealand and 
sign posts stripped black & white in Trinidad.  
5) Orange yellow flags carried by pedestrians crossing the road increases 
pedestrian visibility and conspicuity; and increased intensity of roadway 
lighting.  
6) Crossing paved with coloured material different from roadway pavement 
colour. 
 
Drivers and Pedestrian Reminders 
Crossing signs placed in the roadway reminding road users of their obligations. 
Example is a sign placed along the centre line of the roadway reading “State Law: 
Yield to Pedestrians in Crosswalk” was cited to be helpful to drivers in 
Washington, DC. (Redmon, 2003); A similar sign reading “STOP FOR     “ was 
associated with drivers’ yielding rate  as high as signal controlled crossings 
(Fitzpatrick, et al., 2007); “Look Right” and “Look Left” messages on signs and 
    
  26 
as pavement markings are used in UK to remind pedestrians of the direction to 
look for vehicles before stepping on a crossing (Tan and Zegeer, 1995); Placards 
demonstrating the meaning of pedestrian signal are used in North America 
(Redmon, 2003); Advance signing instructing drivers to yield to pedestrian when 
flashing lights are activated were found to increase driver yielding behaviour at 
intersections (Van Houten et al., 1998). 
 
Reduce Vehicle Speeds 
Reduced vehicle speeds at pedestrian crossings is accomplished by placing the 
crossing between two speed cushions such as humps, placing the crossing on top 
of a speed table—raised crosswalk, providing rumble strips upstream the 
crossing, narrowing the roadway at the crossing by curb extension or middle 
refuge islands, and providing warning flashing lights in the advance zone. 
However, pedestrians in Sweden had a problem of predicting whether the driver 
will stop for them at sites using speed cushions (Leden et al., 2006). 
 
Minimize Vehicle-Pedestrian Conflicts and Pedestrian Exposure 
At a pedestrian crossing, pedestrians are separated from vehicles in time or 
space. Time separation involves traffic signal control with pedestrian phase. 
Vertical spatial separation is accomplished by pedestrian bridge while horizontal 
spatial separation is accomplished by shifting the vehicular stop line further 
upstream of the crossing. Pedestrian barriers between footpaths and roadway are 
used to prohibit pedestrians crossing the roadway at any location except at the 
designated place such as a crosswalk. Provision of a refuge island at the middle 
of the multilane road crosswalk reduces pedestrian exposure by shortening the 
distance a pedestrian is required to cross at a time and help the pedestrian in 
evaluating gaps between traffic streams. Median refuge islands when used with 
Danish offsets, requiring pedestrians to walk facing oncoming traffic, provides 
them a better view of oncoming traffic and allows drivers to clearly see 
pedestrians. They also diminish vehicles overtaking at the crossing. 
    
  27 
Pedestrian Detection 
Activated in-pavement lighting and animated signal have been used to alert the 
driver of the presence of a pedestrian at the crossing. Pedestrian detection is 
accomplished by push button or automatically by sensors. Retting et al. (2003) 
suggested that crossings with pedestrian signals are more effective when 
accompanied with pedestrian activation systems. 
 
2.3.2 Traffic Controlled Mid-block Pedestrian Crossings 
There are two types of traffic controlled crossings at mid-block locations. First, 
vehicular and pedestrian traffic are controlled by traffic lights. Pedestrian signal 
is activated by pedestrian push button or passive detection. In this category 
vehicle control signal, the display sequence is similar to that at signalized 
intersection, i.e., red to green to yellow. Three variations of this type include 
Pelican (Pedestrian Light Controlled), Toucan (Two Can Cross), and Puffin 
(Pedestrian User-Friendly Intelligent). Countdown signal informing pedestrians 
the remaining time for the walk signal is a recent concept more applicable at 
signalised intersections.  
 
Second, vehicular signal uses two red and one yellow signals. It has five display 
indication patterns as opposed to three at standard signalised intersection 
(FHWA, 1988). Fitzpatrick et al. (2006) summarised the signal display sequences 
variations among installations practices. Half signals in the city of Seattle dwell 
in steady green and then cycle to steady yellow and then steady red when 
activated by a pedestrian. The HAWK (high-intensity activated crosswalk) 
signals in Tucson are dark until activated by a pedestrian; then they cycle 
through flashing yellow, steady yellow, steady red, and then flashing red. Half 
signals in the Vancouver dwell in flashing green and, on activation, steady green, 
steady yellow, and then steady red. The mid-block pedestrian signal in Los 
Angeles shows a green arrow, cycles to a steady yellow, and then cycles to 
steady red during the walk interval. During the flashing Don’t Walk interval, 
    
  28 
drivers see a flashing red indication and, after stopping, may proceed if the 
crosswalk is not occupied.  
    
  29 
CHAPTER THREE: METHODOLOGY 
 
Evaluation of the new zebra crossing was done through triangulation method by 
interviewing drivers, interviewing pedestrians, and observing drivers’ yielding 
behaviour in the field. 
 
3.1 Observed Drivers’ Yielding Behaviours 
Drivers’ yielding behaviours at pedestrian crossings were observed in the field. 
Factors that were thought to influence driver’s yielding behaviour were 
accounted for. They include flashing light operation status, vehicle size, vehicle 
use, and incipient pedestrian-vehicle conflict. Conflicts were measured by 
observing approaching drivers’ action when the pedestrian was on the 
crosswalk. 
  
3.1.1 Experimental Design 
To assess the effect of flashing lights on drivers’ yielding behaviours, a cross-
section design with a control group and a treatment group was employed. The 
control group has “priority-pedestrian crossing” under installation. These 
crossings would have reached at a stage whereby all elements including flashing 
lights hardware have been installed but not yet activated. The treatment group 
consisted of similar crossings where flashing lights are activated and are 
working. Selection of crossings in the control group was largely determined by 
the schedule of contractors installing these new crossings with an objective of 
covering wider geographical area as possible. Within the neighbourhood of each 
crossing in the control group, at least one crossing was selected to make up the 
treatment group. All crossings were located on 2-lane, 2-way roads except one 
crossing on a 2-lane, one-way street in urban environment. Subjects were drivers 
at these crossings who encounter a single pedestrian already stepped on the 
crossing area. Only platoon leaders whom braking behaviour is not likely to be 
    
  30 
influenced by other than a pedestrian in the crossing, were observed. It is noted 
that a single vehicle is considered to be a platoon of one vehicle. 
 
3.1.2 Data Collection 
Data was collected in 2005 and 2006 during daylight hours. Two observation 
sessions, one for each approach, were made at each crossing in both control and 
experimental groups. The observer located few meters upstream of the crossing, 
far enough not to distract drivers approaching the crossing, but close enough to 
observe driver’s breaking behaviour as they approach the crossing. In order to 
expedite data collection, a researcher would selected a free flowing vehicle or a 
platoon leader approaching and then begin crossing the roadway. The crossing 
was done taking into consideration safety of the crossing person. This procedure 
was also used by Van Winkle and Neal (2000) who also associated it with test 
consistency. 
 
In response to the presence of the pedestrian in the crossing and an approaching 
vehicle, the second researcher observed the following: 1) driver’s action (stop, 
slow, nothing), 2) vehicle size and registration type, 3) pedestrian’s walking 
direction (into or away from vehicle’s trajectory path). Brake light indication was 
used as a criterion to differentiate between slowing and doing nothing. 
 
Vehicles were classified into three size groups: small, medium, and large. Small 
vehicles include passenger cars and sport utility vehicles (SUVs) while medium 
vehicles included pick-ups, vans, and small trucks up to approximately three 
tons. Large vehicles include trucks above three tons and buses. Vehicle 
deceleration rate depends on the vehicle size, with small vehicles having higher 
rates than bigger vehicles.  
 
Driver’s decision to yield for the pedestrian who is on the crossing is likely to be 
influenced by the potential of vehicle-pedestrian conflict.  DeVeauuse et al. (1999) 
    
  31 
found that drivers comply more with stop sign at pedestrian crosswalks when 
pedestrians are present. If the pedestrian walking direction is likely to cause 
vehicle-pedestrian conflict, the position of the pedestrian was recorded as critical 
region and non-critical region otherwise. Figure 3.1 demonstrates the two 
pedestrian regions at the crossing.  
 
A vehicle in Trinidad is registered either for private use; carrying passengers for 
hire/fare; renting; or transporting goods. Their respective registration number 
plates start with letters P (Private), H (Hire), R (Rent), or T (Transport) 
respectively. Drivers of hire vehicles have been considered to disobey traffic 
regulations more than other groups. 
 
3.1.3 Response Variable 
The response variable was defined as the proportion of drivers yielding to 
pedestrians on the crossing area. The guidelines on the use of priority crossings 
requires driver to accord pedestrian a precedence when the pedestrian is on the 
Pedestrian in this 
area walking in 
either direction 
was considered to 
be in vehicle’s 
path. 
Pedestrian in this area 
is only considered to 
be in vehicle’s path if 
travelling in this 
direction. 
Vehicle travelling 
in this direction 
Figure 3.1: Definition of pedestrian’s region 
    
  32 
crossing. This requires the driver to yield to a pedestrian, but do not specifically 
requires the driver to stop. In this study two levels of driver yielding were set at: 
1) stopping and 2) at least slowing down (stop and slow). The compliment of 
proportion of drivers who at least slows down is the same as the proportion of 
drivers who do nothing to allow pedestrians to cross the road. 
 
3.2 Drivers’ Opinion 
In 2005 a questionnaire survey was carried out to the general motorists to solicit 
drivers’ views on pedestrian crossings and pedestrian safety, and to test drivers’ 
understanding of selected crossing elements. The survey was administered 
through a mail questionnaire. While the major setback of mail questionnaire is 
the difficulty of attaining high response rate, methods that may yield higher 
response rates such as roadside survey or interview could not be used because of 
large number of questions and limited resources.  
 
The island of Trinidad was stratified in three geographical regions of north, 
central, and south. A total of 1000 questionnaires were distributed to drivers at 
entrances of the shopping centres. The same number of questionnaires was 
distributed to each of the three regions. It was expected that majority of shoppers 
at these centres would constitute local residents. A self-addressed, business reply 
envelope was enclosed within the questionnaire package to enable respondent 
mail back the questionnaire at no postage cost on the part of the respondent. 
 
3.2.1 The Questionnaire 
A one-page questionnaire consisted of 10 short multiple-choice questions. The 
intent of multiple-choice questions was to offer convenience on the part of the 
respondent and easy of analysis on the part the research team.  The questionnaire 
solicited information from the driver in three areas of 1) respondent’s 
demographic factors; 2) driver’s understanding and yielding behaviour at zebra 
    
  33 
crossings, and 3) estimate of pedestrian crossing problem in the country. A 
sample questionnaire is shown in the Appendix. 
 
3.2.2 Demographic Factors 
Demographic factors sought were respondent’s gender, age, and period of 
driving experience. Drivers with longer experience have learned from the past 
the driving environment, and therefore they are expected to be safer drivers. 
Young drivers especially teen drivers, on the other hand, they don’t only lack 
sufficient knowledge and driving experience, but they are also known for their 
high-risk taking behaviour that may degrade road safety. High crash rates 
involving younger drivers reflect the effect of age, since older drivers who are 
similarly inexperienced has somewhat lower crash rates (Cooper et al., 1995). 
Males have were found higher risk takers than female drivers (Hemenway and 
Solnick, 1993), with men being more likely than women to report having driven 
after drinking, regularly exceeding the speed limit, and running a red light.  
 
3.2.3 Drivers’ Understanding of Crossing Elements 
Traffic Control Devices (TCDs) are designed by engineers to be used by road 
users (drivers, pedestrians, cyclists, etc.). Sometimes road users understand 
TCDs differently from the intended meaning (Stokes et al., 1996; Ford and Picha, 
2000). It is expected that drivers’ understanding correctly the meaning of a TCD 
would be a prerequisite for its effectiveness. Some standards like the Manual on 
Uniform Control Devices (MUTCD) identify “conveyance of a clear, simple 
meaning” as one of the five requirements that should be met by a TCD (FHWA, 
1988). 
 
Three questions were included in the questionnaire to test the drivers’ 
understanding of three main elements of a priority-pedestrian zebra crossing. 
The three elements are: - 1) zebra markings, 2) zigzag lines, and 3) flashing lights. 
A coloured photograph of the crossing was part of the questionnaire for easy 
    
  34 
reference to these crossing elements. Respondents were also asked whether their 
understanding of the meaning of zebra markings is based on their understanding 
of traffic laws or/and through driving experience.  
 
3.2.4 Self-Reported Yielding Behaviour 
Apart from field observations, an alternative method of measuring drivers’ 
yielding behaviour is to ask drivers themselves—self-reported behaviour. While 
the self-reporting method could be less costly, it suffers the self-reporting bias -- 
when the respondent tries to impress the interviewer thereby hiding bad 
behaviours. Questions were worded so as to minimize the magnitude of self-
reporting bias, since it was though that such magnitude depends on the method 
of interview used as well as wording of the question. Drivers were asked to 
indicate their responsibility at zebra crossings when a pedestrian is waiting to 
cross the roadway. We avoided asking directly what they do in order to 
minimize self-reporting bias. 
 
3.2.5 Estimate of Pedestrian Crossing Problem 
Four questions intended to estimate subjectively the magnitude of the pedestrian 
safety problem in the country. In these questions respondents were asked:- 
• How often do they not yield to pedestrians;  
• The frequency the respondent sees other drivers not yielding to pedestrians;  
• If they know any pedestrian who was a victim of “crossing the road” 
phenomena. 
 
3.2.6 Effectiveness of the New Zebra Crossing 
Public and stakeholders’ support for any project is essential in sustaining such 
project for long term. In this case we defined the effectiveness of the new 
crossing in terms of its ultimate goal of reducing pedestrian related crashes. 
Drivers were asked to give their opinions on whether installation of new 
pedestrian crossings would help in decreasing pedestrian related accidents. 
    
  35 
 
3.3 Field Visits and Pedestrian Survey 
Field visit to priority-pedestrian zebra crossings was intended to gain personal 
experience about the crossings and their usage. The survey was done through 
personal interviews to some pedestrians who used the crossings. They include 
school principals and pedestrian passing-by the crossings that showed interest, 
at the time the researchers were observing drivers’ yield behaviours. Principals 
of schools that were in the vicinity of some of these crossings were interviewed. 
They were considered to represent views of school children who use those 
crossings. 
    
  36 
CHAPTER FOUR: ANALYSIS AND DISCUSSIONS 
 
Data were summarised and analysed using Statistical Package for Social Sciences 
(SPSS) software. Discussion of the results was based on the results of the 
analyses, together with authors’ experience of pedestrian safety in Trinidad. 
Statistical differences between factors of interest were tested at 95 percent 
significance level. 
 
4.1 Observed Drivers’ Yielding Behaviour 
A total of 1630 drivers were observed at 22 crossings. Nine hundred and nine 
drivers were observed at fourteen crossings in the treatment group (flashing 
lights-on), while seven hundred twenty one drivers were observed at nine 
crossings in the control group (flashing lights-off). Only one crossing was 
observed in both conditions of lights-on and -off. Table 4.1 summarises the 
crossings’ information (name, number of observations, flashing light operation 
condition, and the year it was observed). To assess the effect of flashing lights 
operation on yielding behaviour of drivers, analysis was done either using one-
tail t-test or chi-square (?
2
) test on drivers yielding behaviour. The effects of 
pedestrian walking direction, vehicle size and vehicle registration type were 
accounted for. Figure 4.1 indicates that, on the overall, 14 percent of drivers at all 
crossings stopped for pedestrians, while 66 percent slowed down to allow 
pedestrians cross the roadway, and 20 percent did not yield at all. 
 
4.1.1 Effect of Flashing Lights Operation 
Table 4.2 and Figure 4.1 shows that the proportion of drivers who stopped for 
pedestrians at crossings where flashing lights were activated, i.e., 22%, is five 
times greater than at crossings where flashing lights were not activated, i.e., four 
percent. The difference is statistically significant (p-value < 0.001). Similarly the 
corresponding proportions for drivers who stopped or slowed down to allow the 
pedestrian to cross the roadway were 83% and 75% respectively; the difference 
    
  37 
being statistically significant (p-value <0.001). What could be concluded from 
these results is that flashing lights did increase the proportion of drivers who 
stopped and those who at least slowed down to allow pedestrians to cross the 
road. Since the proportion of drivers who at least slowed down is a compliment 
of those who did nothing, it is concluded that flashing lights reduced the 
proportion of drivers that neither stopped nor slowed down to allow pedestrians 
to cross the road. 
 
Table 4.1: Summary of drivers’ observed yielding behaviour data 
Control Group: Flashing Lights Off Treatment Group: Flashing Lights On 
Site Name Observations Site Name Observations 
Bossierre Village RC (2006) 80 3
rd
 St. & Access Road, Barataria 
(2005) 
18 
Monroe Rd. (2006) 80 3
rd
 St. & PBR, Barataria (2005) 61 
Maraval RC Sch. (2006) 80 BWIA HQ, Piarco (2005) 77 
Chaguanas Govt. Primary 
school (2006) 
81 Enterp. Govt. School--
Homeland (2006) 
80 
Orange Valley & Bay Rd. 
(2006) 
80 Preysal Hindu School, Couva 
(2006) 
80 
San Fdo Boys R.C (2006) 80 Prison Golden Gr. Rd (2005) 33 
Servila Private Primary 
School. (2006) 
80 WMR West Moorings (2006) 80 
SMR Curepe (2006) 80 SMR Curepe (2006) 80 
SMR la Paille-Caroni 
(2006) 
80 Warren ASJA-SMR (2006) 80 
  Bryn Maur, Petit Valley (2006) 80 
  Hindu sch., Freeport (2006)  80 
  Marabella Sen. Comp. (2006) 80 
  Charl. Presby (2006) 80 
Total 721 Total 909 
 
Table 4.2: Effect of flashing lights on driver's yielding behavior 
Flashing 
Lights Status 
% Stopping p-value % Stopping or Slowing p-value 
ON (n=909) 22.1 83.5 
OFF (n=721) 4.4 
<0.001 
75.1 
<0.001 
    
  38 
 
Figure 4.1: Effect of flashing lights on driver's yielding behaviour
22
61
17
4
71
25
14
66
20
0
10
20
30
40
50
60
70
80
Stop Slow Nothing
Driver's yielding
Pe
r
c
e
n
ta
ge
Lights ON
Lights OFF
Total
 
 
4.1.2 Effect of Pedestrian’s Walking Direction 
In this section we are trying to answer the question of whether flashing lights 
effectiveness depends on the pedestrian walking direction with reference to 
driver/vehicle’s projected path. First we investigate if pedestrian’s walking 
direction affects drivers’ yielding behaviour at all. Table 4.3 shows that 23.6%, of 
drivers stopping for pedestrians when a pedestrian was walking towards 
vehicle’s path, is about four times higher than the corresponding proportion, i.e., 
6.4% when a pedestrian was walking away from vehicle’s path. The difference 
between the two proportions is statically significant (p-value<0.001). Similarly 
the proportion (84.4%) of drivers who at least slowed down to allow a pedestrian 
to cross the road when the pedestrian is walking towards vehicle’s path is higher 
than the corresponding proportion (75.9%) when a pedestrian is walking away 
from the vehicle’s path. The difference between the two proportions is also 
    
  39 
statistically significant (p-value<0.001). It can be seen that more drivers yielded 
to pedestrians when incipient conflict was perceived, i.e., pedestrian walking 
towards the direction of the vehicle’s path. 
 
Table 4.3: Effect of pedestrian’s walking direction on driver’s yielding 
behaviour 
Pedestrian’s Direction % Stopping p-value 
% Stopping 
or Slowing 
p-value 
Towards vehicle’s path (n=751) 23.6 84.4 
Away from vehicle’s path (n=879) 6.4 
<0.001 
75.9 
<0.001 
  
The effect of flashing lights on driver’s yielding behaviour taking into 
consideration of pedestrian walking direction was also examined and is shown 
in Table 4.4. For the case of a pedestrian walking towards vehicle’s path, the 
proportions of drivers who stopped for pedestrians at crossings with flashing 
lights activated (37.4%) was higher than at crossings with flashing lights off 
(5.3%). The difference is statistically significant (p-value < 0.001). Similar pattern 
is true for drivers who at least slowed down in which 87.6% and 80.2% at least 
slowed down for activated and inactivated flashing lights respectively.  
 
For pedestrian walking away from driver’s path, proportions of yielding drivers 
were lower but with the same pattern as with pedestrian walking towards 
driver’s path situation, as depicted in Table 4.4. It could be concluded that 
flashing lights did increased the proportion of drivers stopping, and reduced the 
proportion of drivers that neither stopped nor slowed down to allow pedestrians 
to cross the road, irrespective of the pedestrian’s direction. 
 
    
  40 
 
Table 4.4: Effect of flashing lights and pedestrian’s walking direction on 
driver’s yielding behaviour 
Flashing 
Lights Status 
% Stopping p-value % Stopping or Slowing p-value 
Pedestrian walking towards vehicle’s path 
ON (n=428) 37.4 87.6 
OFF (n=323) 5.3 
<0.001 
80.2 
<0.003 
Pedestrian walking away from  vehicle’s path 
ON (n=481) 8.5 79.8 
OFF (n=398) 3.8 
<0.002 
71.1 
<0.001 
 
4.1.3 Effect of Vehicle Size on Drivers’ Yielding Behaviour 
Larger vehicles may require longer braking distances than smaller vehicles. It is 
therefore expected that larger vehicles will have difficulties of stopping for 
pedestrians than smaller vehicles. Table 4.5 indicates a trend of smaller vehicles 
stopping for pedestrians more than larger vehicles, and in turn larger vehicles 
slowing down than smaller ones. The differences are statistically significant. 
These patterns may suggest the differences between vehicle sizes in bringing 
them to a stop condition as expected. There also exist a trend of larger vehicles 
yielding (at least slowing down) more than smaller vehicles—it could be that 
drivers of smaller vehicles who don’t yield often are over-confident in their 
vehicle control capability thinking that they can manoeuvre around such hazards 
without even slowing down.  
 
Table 4.5: Effect of vehicle size on drivers’ yielding behaviour 
Vehicle size % stopping p-value 
% stopping 
or slowing 
p-value 
Small (n=1189) 15.6 76.4 
Medium (n=364) 12.1 81.0 
Large (n=77) 5.2 
<0.016 (?
2
-test) 
88.3 
<0.014 (?
2
-test) 
 
The effect of flashing lights on driver’s yielding behaviour in each vehicle group 
is shown in Table 4.6. In all six cases of vehicle size and yielding criterion 
combination, proportion of drivers who yielded to pedestrians at crossings with 
    
  41 
flashing lights-on was higher than corresponding proportion at crossings with 
flashing lights-off. The differences are statistically significant except in one case 
of large vehicles with at least slowing condition. It could be concluded that 
generally flashing lights increased drivers yielding for both small and large 
vehicle drivers. Difference between lights-on and lights-off  across all vehicle 
groups in Table 4.6 shows that flashing lights were more effective in influencing 
driver to stop than yielding (at least slowing down). 
 
Table 4.6: Effect of flashing lights within vehicle size on drivers’ yielding 
behaviour 
Flashing 
Lights 
Small vehicles p-value 
Medium 
vehicles 
p-
value 
Large 
vehicles 
p-
value
Stop condition 
ON 24.4% (n=664) 17.2% (n=203) 95%(n=42) 
OFF 4.4% (n=525) 
<0.001 
5.6% (n=161) 
<0.001 
0%(n=35) 
<0.03 
Stop or Slow conditions 
ON 82.5% (n=664) 85.2% (n=203) 90.5%(n=42) 
OFF 74.3% (n=525) 
<0.001 
75.8% (n=161) 
<0.01 
85.7%(n=35) 
<0.26 
 
4.1.4 Effect of Vehicle Use on Drivers’ Yielding Behaviour 
Experience indicates that drivers of “hire” vehicles are more aggressive thereby 
violating traffic laws more often. We want to find out if flashing lights had the 
similar effect for “hire” and “non-hire” vehicle drivers. Comparison of 
proportions of drivers who yielded to pedestrians between “hire” and “non-
hire” vehicles in Table 4.7 & 4.8 did not include “large” vehicles because all 
“hire” vehicles were observed in “small” and “medium” vehicle groups. 
Proportion of drivers in “hire” group was slightly higher than in corresponding 
“non-hire” group for both yielding criteria of stopping and at least slowing down 
(Table 4.7). However, these differences are not statistically significant. It can be 
concluded that type of vehicle use had little impact on driver’s yielding 
behaviour. 
 
    
  42 
Table 4.7: Effect of vehicle use on driver’s yielding behaviour 
Vehicle Use % Stopping p-value % Stopping or slowing p-value 
For Hire (n=582) 15.3 78.5 
Non-hire (n=971) 14.4 
<0.319 
79.9 
<0.255 
 
Comparisons between flashing lights conditions by vehicle use are sown in Table 
4.8. In all four cases the proportion of drivers yielded to pedestrians is higher at 
crossings with flashing lights-on than at crossings with flashing lights-off. The 
differences are statistically significant. It could be concluded that flashing lights 
did increase driver yielding in both “hire” and “non-hire” vehicle groups. 
 
Flashing lights seem to have bigger effect for stopping condition. Proportions of 
stopped drivers at flashing lights-on and flashing lights–off in “hire” group was 
22.9% and 5.2% respectively. The difference is 17.7 percentage points. The 
corresponding values in “non-hire” group are 22.6%, 4.4%, and 18.2 percentage 
points respectively. On the other hand, proportion of drivers that yielded (at 
least slowed) in “hire” group at flashing lights-on was 85.2% compared with 
69.6% resulting into a difference of 15.6 percentage points. Corresponding values 
in “non-hire” group are 81.9% and 77.5% respectively, resulting into a difference 
of only 4.4 percentage points. 
 
Table 4.8: Effect of flashing lights by vehicle use on driver’s yielding 
behaviour 
Flashing Lights % Stopping p-value % Stopping or slowing p-value 
Vehicles for hire 
ON (n=332) 22.9 85.2 
OFF(n=250) 5.2 
<0.001 
69.6 
<0.001 
Vehicles not for hire 
ON (n=535) 22.6 81.9 
OFF(n=436) 4.4 
<0.001 
77.5 
<0.046 
  
 
 
    
  43 
4.1.5 Modelling of Driver’s Yielding Behaviour 
While the analysis in previous sections looked at driver yielding factors 
(independent variables) one at a time, i.e. univariate analysis. In this section these 
factors are analysed together, i.e. multivariate analysis, in order to increase the 
statistical power of the analysis. Four issues that will be looked at as related to 
statistical modelling of driver’s yielding behaviour at zebra crossing in a 
multivariate context are:- 
• Research question(s), 
• Approach and methodology, 
• Model specification, and 
• Results & interpretations. 
 
Research Questions 
Questions addressed include: what is the probability of a driver yielding to a 
pedestrian in the zebra crossing given set of conditions of attributing factors 
(predicting equation)? Does flashing lights operation, pedestrian region, vehicle 
size, or vehicle registration type make difference in driver’s yielding behaviour 
prediction?  
 
Approach and Methodology 
To answer the above questions, we conducted logistic regression analysis with 
one dichotomous dependent variable (driver yielding) and four independent 
variables of flashing lights, pedestrian region, vehicle use, and vehicle size. 
Logistic regression analysis was preferred over other alternatives because of its 
flexibility, and the dichotomous outcome of the dependent variable. It has no 
assumptions about the distributions of the independent variables; the 
independent variables have not to be discrete, as it can accommodates a mixtures 
of discrete and continuous variables; it cannot produce negative predicted 
probabilities. 
    
  44 
Model Specification 
The model comprises four independent variables (covariates) and one binary or 
dichotomous dependent variable of driver yield (yes/no). Yielding condition 
defined as at least slow down condition. The four covariates were two levels 
each. They are: 1) flashing light operation (on/off), 2) pedestrian region 
(towards/away from vehicle’s path), 3) vehicle use (hire/non-hire), and 4) 
vehicle size (passenger car/non-passenger car). 
 
Results and Interpretations 
The constant only model indicates that 79.8% of drivers were likely to yield to 
pedestrians. Incorporating independent variables in the model (full model) was 
justified at p-value less than 0.05. Of the four independent variables added, 
vehicle size and vehicle use were not significant in the model (p>0.05). The 
reduced model retained only two variables of flashing light operation & 
pedestrian region. Figure 4.2 shows an extract of SPSS output for modelling 
drivers’ yielding behaviour. The probability (p) that a driver will yield to a 
pedestrian on the crossing is evaluated using equation 4.1.  
)(
1
1
pLogit
e
p
?
+
= .......................................................................................................(4.1) 
Where: Logit (p) { =1.183+0.537X
1
+0.506X
2
-0.313X
3
-0.046X
4 
 for full model; 
 =0.885+0.538X
1
+0.507X
2 
 for reduced model}. 
X
1
 = Pedestrian walking direction { =1 if pedestrian is walking towards 
vehicle’s trajectory path;  
 = 0 Otherwise}.  
X
2
 = Flashing lights operation        { =1 if activated;  
 = 0 Otherwise}.  
X
3
 = Vehicle registration type        { =1 If for hire;  
 = 0 Otherwise}.  
X
4
 = Vehicle size                                { =1 if passenger car;  
 = 0 Otherwise}.  
  e ? 2.718.  
    
  45 
Pedestrian walking direction was found to influence driver yielding than 
flashing light operation factor. It has an odds ratio of 1.712 which is interpreted 
to mean that the odds that a driver will yield to a pedestrian walking towards 
vehicles trajectory path is 1.712 times the odds that  when the pedestrian is 
walking away from vehicle’s trajectory path, with other factors held constant. It 
is more likely to yield to a pedestrian when the pedestrian is walking towards 
vehicle path than otherwise. Flashing lights operation has odds ratio of 1.66 
meaning that the odds for a driver yielding to a pedestrian when flashing lights 
are operational is 1.66 times the odds when flashing lights are not activated, 
other factors held constant. In other words: drivers are likely to yield to 
pedestrian when flashing lights are activated than otherwise. 
 
Full Model B S.E. Wald df Sig. Exp(B) 
Step 1(a) lights .506 .125 16.369 1 .000 1.658 
  region .537 .129 17.381 1 .000 1.711 
  hire -.313 .257 1.481 1 .224 .731 
  veh_size -.046 .211 .047 1 .828 .955 
  Constant 1.183 .181 42.493 1 .000 3.264 
Reduced Model B S.E. Wald df Sig. Exp(B) 
Step 1(a) lights .506 .125 16.369 1 .000 1.658 
  region .537 .129 17.381 1 .000 1.711 
  hire -.313 .257 1.481 1 .224 .731 
  veh_size -.046 .211 .047 1 .828 .955 
  Constant 1.183 .181 42.493 1 .000 3.264 
Figure 4.2: Extract of SPSS output of modelling drivers’ yielding behaviour 
 
Probabilities Estimation 
Probabilities (Pr(y)) for a driver yielding to a pedestrian under four different 
cases as computed from reduced model of equations 4.1 are:-. 
Flashing lights off, pedestrian walking away from vehicle’s path: 
71.0
1
1
)Pr(
885.0
=
+
=
?
e
y  
Flashing lights off, pedestrian walking toward vehicle’s path: 
    
  46 
80.0
1
1
)Pr(
)538.0885.0(
=
+
=
+?
e
y  
Flashing lights on, pedestrian walking away from vehicle’s path: 
81.0
1
1
)Pr(
)507.0885.0(
=
+
=
+?
e
y  
Flashing lights on, pedestrian walking toward vehicle’s path: 
87.0
1
1
)Pr(
)507.0538.0885.0(
=
+
=
++?
e
y  
 
4.1.6 Comparison with Other Studies 
Comparison of drivers’ yielding rates from this study with other studies is an 
attempt to gauge the level of success of flashing lights with other comparable 
devices elsewhere. Comparison is limited to improved mid-block zebra crossings 
where vehicular traffic is not controlled by traffic lights, and there is no 
pedestrian activation. Table 4.9 summarizes the comparison. Drivers’ yielding 
rates observed in Trinidad are within ranges of similar studies elsewhere despite 
differences in study methodologies and roadway environment.
    
  47 
Table 4.9: Comparison with other studies on drivers’ yield rate 
Study Characteristics 
Without 
Improvement 
With 
Improvement 
Dependent Variable: Driver Stopping or Slowing 
Trinidad (current study) 
23 mid block zebra crossing; cross-sectional study; 
supplemented with flashing lights 
75% 84% 
Central Avenue, St. 
Petersburg, FL. (Van Houten, 
and Malenfant, 2001) 
Mid block crosswalk on multi lane roadway; before-and-after 
study; treated with animated eye signal. 
15% 62% 
Salt Lake city in Utah and 
Kirkland in Washington 
(Fitzpatrick et al., 2006) 
Cross-sectional study on four crossings, using pedestrian 
flags, at 2-6 through-lane, high-volume, high-speed 
roadways. 
Not available 74% 
Redmond, Washington 
(Fitzpatrick et al., 2006) 
Cross-sectional study on three crossings, using in-street 
crossing sign, at 2-3 through-lane, high-volume, high-speed 
roadways. 
Not available 90% 
 Tucson, Arizona and Austin, 
Texas (Fitzpatrick et al., 2006) 
Cross-sectional study on two crossings, using high visibility 
signs and markings, at 4 through-lane, high-volume, high-
speed roadways. 
Not available 20% 
 Tucson, Arizona  
(Fitzpatrick et al., 2006) 
Cross-sectional study on one crossings, using high visibility 
signs and markings, at 4 through-lane, high-volume, high-
speed roadway. 
Not available 91% 
Wilcox Blvd. Chattanooga, 
Tennessee 
Before-and-after on one mid-block crossing with actuated 
flashing lights at two locations in advance zone and at 
crossing location (Van Winkle and Neal, 2000) 
48% 70-81% 
Dependent Variable: Driver Stopping 
Trinidad (current study) 
23 mid block zebra crossing, cross section study; 
supplemented with flashing lights 
4% 22% 
North Carolina (Clark et al., 
1996) 
11 un-signalized mid block crossings; cross-sectional study; 
treated by replacement of existing signs with fluorescent 
strong yellow-green 
6% 10% 
Wilcox Blvd. Chattanooga, 
Tennessee 
Before-and-after on one mid-block crossing with actuated 
flashing lights at two locations in advance zone and at 
crossing location (Van Winkle and Neal, 2000) 
9% 42-50% 
    
  48 
4.2 Drivers’ Opinion 
Out of 1000 questionnaires, one hundred and ninety were returned (19% 
response rate). Fifty-five percent of respondents were females. Fifty-eight percent 
of returned questionnaires came from south region, 39 percent from central 
region, and 13 percent from north region. Table 4.10 gives descriptive statistics of 
the respondents’ age and licensed period to drive. Table 4.11 shows that age and 
driving experience are highly correlated.  
 
Table 4.10: Descriptive statistics for survey respondents 
Parameter N Minimum Maximum Mean Std. Deviation   
Age 180 17 yrs 75 yrs 38.6 yrs 14.3 yrs   
Driving Experience 174 1 yrs 55 yrs 17.3 yrs 12.7 yrs   
  
Table 4.11: Correlation between respondents’ demographic factors 
Demographic factors 
Measure 
Age vs Experience Age vs Gender Gender vs Experience 
Pearson 0.914
a
 -0.107 -0.161 
Kendall’s tau-b 0.777
a
 -0.062 -0.109 
Spearman’s rho 0.906
a
 -0.075 -0.131 
a
 Correlation is significant at the 0.01 level (2-tailed) 
 
4.2.1 The Meaning of Zebra Markings 
One hundred and eighty-seven responded to the question on the meaning of 
zebra markings from the drivers’ point of view. Gender, age, and driving 
experience were not statistically significant different on this issue. Distribution of 
responses is summarised as part of Table 4.12 from which the following 
observations are made:- 
1) The proportion of drivers who understand that zebra markings means 
“stopping” is higher than for those who understand it to mean “slowing 
down”, i.e., 60% and 9% respectively. This pattern is consistent in both cases 
of when the “pedestrian is about to enter the crosswalk” and when the 
“pedestrian is in the crosswalk”, i.e., 23% and 5% respectively; 
    
  49 
2) The presence of respondents, about 3%, who understand zebra markings to 
mean nothing to the driver, is a matter of concern. Zebra markings at 
pedestrian crossings are among the oldest traffic control devices used 
worldwide to facilitate pedestrians cross roads legally and safely. It would be 
expected that all drivers would at least understand that these markings mean 
something to the driver. 
 
Logistic regression was performed to measure the contribution of gender, age, 
and driving experience on respondents’ understanding the meaning of the zebra 
markings. First the dependent variable was collapsed from five responses to only 
two (correct and incorrect). One hundred and sixty-three cases were used in the 
analysis, which satisfies the requirement of 15 to 20 cases per independent 
variable. Table 4.13 is an extract from SPSS’s output from which it is concluded 
that neither gender nor age nor experience were significant (p>0.05) in predicting 
driver’s understanding of the meaning of zebra markings. 
4.2.2 The Meaning of Flashing Lights 
Yellow flashing lights at priority-pedestrian zebra crossings are intended to 
indicate the presence of a pedestrian crossing so that a driver can prepare to 
yield in case a pedestrian is using the crossing. Inclusion of “prepare to stop” as 
one of optional answers in the questionnaire was to translate the meaning of 
flashing light to the respondent. Summary of the result is given in Table 4.12, 
from which it is noted that 12% of respondents misunderstood the meaning of 
these lights. Before the introduction of the new zebra crossing in Trinidad, 
yellow flashing lights in road traffic environment was only used at signalised 
intersection, when signals are operating in a safe mode. Such experience could 
have contributed towards misunderstanding of flashing lights as used at 
pedestrian crossings. 
 
    
  50 
Logistic regression was performed to measure the contribution of gender, age, 
and driving experience on respondents’ understanding the meaning of the 
flashing lights. First the dependent variable was collapsed from five responses to 
only two (correct and incorrect). One hundred and sixty-four cases were used in 
the analysis, which satisfies the requirement of 15 to 20 cases per independent 
variable. Table 4.14 is an extract from SPSS’s output from which it is concluded 
that neither gender nor age or experience were significant (p>0.05) in predicting 
driver’s understanding of the flashing lights. 
 
Table 4.12: Understanding of priority-pedestrian zebra crossing elements 
Response Category Percentage 
Meaning of Zebra Marking to the Driver (N=187) 
Stop for a pedestrian who is about to enter the crosswalk 60 
Slow down for a pedestrian who is about to enter the crosswalk 9 
Stop for a pedestrian who is in the crosswalk 23 
Slow down for a pedestrian who is in the crosswalk 5 
Mean nothing to the driver 3 
Meaning of Zigzag Lines (N=176) 
Slow down 34 
Proceed with caution 59 
No parking 1 
Stop 3 
None 3 
Meaning of Flashing Yellow (N=187) 
Prepare to stop 88 
Changing green to red 3 
Only for Pedestrians 1 
None 8 
 
Table 4.13: SPSS output on respondent attributes’ significance in predicting 
their understanding of the meaning of zebra markings. 
 Factor Estimate S.E. Wald df Sig.(p-value) 
Gender -.171 .378 .203 1 .652 
Age .057 .031 3.294 1 .070 
 Experience -.013 .042 .097 1 .756 
 
  
    
  51 
Table 4.14: SPSS output on respondent attributes’ significance in predicting 
their understanding of the meaning of flashing lights 
 Factor Estimate S.E. Wald df Sig.(p-value) 
Gender .398 .517 .593 1 .441 
 Age .027 .051 .284 1 .594 
 Experience -.013 .056 .050 1 .824 
 
4.2.3 The Meaning of Zigzag Lines 
The meaning of the zigzag lines was the most misunderstood element of the 
crossing. Only one percent identified the correct answer of “no-parking” (Table 
4.12). Logistic regression was performed to measure the contribution of gender, 
age, and driving experience on respondents’ understanding of the zigzag lines. 
First the dependent variable was collapsed from five responses to only two 
(correct and incorrect). One hundred and fifty-four cases were used in the 
analysis, which satisfies the requirement of 15 to 20 cases per independent 
variable. Table 4.15 shows the results of the analysis from which it is can be 
concluded that neither gender nor age or experience were significant in 
predicting driver’s understanding of the zigzag lines. 
 
Table 4.15: SPSS output on respondent attributes’ significance in predicting 
their understanding of the meaning of zigzag lines 
 Factor Estimate S.E. Wald df Sig.(p-value) 
Age -.302 .344 .769 1 .380 
Gender 17.748 4511.555 .000 1 .997 
 Experience .309 .371 .692 1 .405 
 
Literature indicates that the meaning of zigzag lines has been illusive to others, 
locally as well as internationally. Misunderstanding of zigzag lines was noted in 
a T&T local newspaper article in which the reporter understood zigzag lines to 
mean, “warn motorists to slow down and exercise caution” (Peters, 2004). 
Coincidentally over 90% of respondents in this survey understand zigzag lines to 
mean one of these two meanings given in the newspaper. It is likely that the 
    
  52 
newspaper article had an impact on public understanding of these lines. The 
newspaper article appears to have sourced the information from the government 
and some readers may easily conclude that this reflect understanding of some 
officials within the government.  
 
Thirty-five percent of respondents in the Australia study gave the correct answer 
to the meaning of zigzag lines (Australian Road Research Board, 1989). This 
prompted reconsideration of use of zigzag lines at zebra crossings. Queensland 
prohibited the use of zigzag lines in 2002 because it was not a recognized 
standard and was considered a potential source of confusing motorists, thereby 
increasing the risk to pedestrians and motorists (Queensland, 2002).  
 
Referring to application of zigzag lines at zebra crossing in Great Britain, Tan 
and Zegeer (1995) as well as Davies (1999) understood zigzag lines also to 
provide a cue to oncoming motorists that they are approaching a crosswalk, 
while others understand the lines to prohibit parking and overtaking only 
(Inman and Davis, 2007; MTC, 2007). In any event, for zigzag lines to provide a 
significant element of advance warning to the crossing more than zebra 
markings, they will have to extend a sufficient distance upstream the crossing 
such as the application in Australia (Figure 4.3). UK Highway Code 191 explicitly 
prohibits parking at zigzag lines on the crossing, and prohibits overtaking at all 
pedestrian crossings (Department of Transport, 2007). 
    
  53 
 
In Trinidad, zigzag lines’ length and their locations relative to crossing are not 
likely to add a significant element of advance warning. In Trinidad, parking 
restriction by zigzag pavement markings is used only at priority-pedestrian 
zebra crossings—it is likely the public is not familiar with the use of zigzag lines 
at pedestrian crossings, because at other locations parking prohibitions signs are 
used instead.  
4.2.4 Awareness of Traffic Laws at Zebra Crossings 
One hundred and eighty-five respondents indicated the source of understanding 
zebra markings. Fifty percent of respondents indicated that their understanding 
of zebra crossing markings comes from traffic laws alone. Twenty six percent 
thinks their understanding is a result of their driving experience or intuition 
alone, while the remaining 24 percent thinks it is a result of both traffic laws and 
experience. This indicate that majority of drivers understands that pedestrians 
are protected to some extent by existing traffic laws. 
Figure 4.3: Zigzag markings used for advance warning  
    
  54 
4.2.5 Drivers’ Yielding Behaviour 
In different ways five questions solicited information about driver’s yielding 
behaviour in a self-reported manner. The first question asked respondents the 
meaning of zebra markings in light of yielding actions. Analysis of responses to 
this question was discussed in subsection 4.2.1 in light of zebra markings 
meaning. In this section we look at the same question in light of drivers’ yielding 
behaviour. Assuming that drivers are likely to act according to their 
understanding of the meaning of a traffic control device, about 97% of 
respondents are likely to yield to pedestrians (Table 4.12). The majority reported 
that they would yield to pedestrian as early as when the pedestrian is about to 
enter the crosswalk. However, one respondent expressed the difficulty of 
understanding the intention of a pedestrian standing at the side of the road. 
 
The second question asked what respondents thinks a driver is required to do for 
a pedestrian who is waiting to cross the road at a marked crossing. 
Overwhelming majority (99%) indicated that they would at least slow down for a 
pedestrian, 92 percent indicating they would stop for a pedestrian (Table 4.16). 
Logistic regression was performed to assess the predictive power of gender, age, 
and driving experience on drivers’ understanding of this issue. Only one of the 
two respondents under “nothing” category gave all information on age, gender, 
and experience. This category was left out from the dependent variable leaving 
only “slow down” and “stop” categories. The number of cases used in this 
analysis was 165. As shown in Table 4.17 all three factors of age, gender, and 
driving experience were not significant (p-values of 0.613, 0.731, and 0.232 
respectively) in predicting driver’s response. 
 
Table 4.16: Driver’s requirement for a pedestrian waiting to cross (n=188)  
Response Category Percentage 
Nothing 1 
Slow down 7 
Stop 92 
    
  55 
  
Table 4.17:  SPSS output on respondent attributes’ significance in predicting 
their requirement for a pedestrian waiting to cross 
 Factor Estimate S.E. Wald df Sig.(p-value) 
Gender .229 .666 .118 1 .731 
 Age .024 .048 .256 1 .613 
 Experience -.068 .057 1.428 1 .232 
 
The third question asked respondents how often they yield for pedestrians at a 
pedestrian crossing. The fourth question asked how often they see other drivers 
yielding for pedestrians at pedestrian crossings. Results of these two questions 
are reported in Table 4.18. Age, driving experience, and gender factors were not 
statistically significant. Few respondents (about 8%) indicated them never or saw 
fellow drivers stopping for pedestrians. Difference between the proportions of 
respondents who “always” yield for pedestrians who “always” see fellow 
drivers yielding for pedestrians was statistically significant. The same is true for 
those who “sometimes” yield for pedestrians. 
 
Table 4.18: Reported frequency on drivers’ yielding for pedestrians at zebra 
crossings 
Driver’s frequency for stopping for pedestrians 
Driver category 
Always
1
 Sometimes
1
 Never
2
 
Reporting of fellow drivers (N=187) 29% 67% 4% 
Self reported (N=188) 55% 41% 4% 
1
 Statistically significant; 
2
 Not statistically significant 
 
The fifth question asked respondents what is the reason when they don’t yield 
for pedestrians. Results are summarised in Table 4.19. Approximately three-
quarters (73%) of 160 respondents indicated that late detection of pedestrian in a 
crosswalk is the major reason for failure to yield. Late detection could be 
attributed to distracted drivers or poor crossing & pedestrian visibility. Also it 
could be used as an excuse by drivers who don’t want to reveal their unsafe 
driving practice (self-reporting bias).  
 
    
  56 
Eight percent don’t yield because they think that pedestrians take long time to 
cross roadways. These are likely to be drivers who are in hurry. Pedestrians 
taking long time to cross and late detection of pedestrian by drivers were also 
reported by Redmon (2003) in which drivers who reported late detection 
indicated sight obstruction and night environment as main reasons. Sixteen 
percent reported that slowing down is enough while three percent expects 
pedestrian to yield to motorists. Motorists think that roads are exclusively for 
vehicles, and therefore yielding to pedestrian is considered a courtesy.  
 
“Pedestrian take long time to cross the road” and “expect pedestrian to yield to 
motorists” may represent aggressive driving habit. Some studies have reported 
significant differences between male and female drivers with regard to 
distraction and aggressive driving behaviour (Shinar and Compton, 2004; Lesch 
and Hancock, 2004). In this study however, there was no significant difference 
between genders. 
 
Table 4.19: Reasons why drivers not yielding for pedestrians at zebra crossings  
Reason (N=160) Percentage 
Don’t see the pedestrian at enough time 73 
Slowing is enough for pedestrian to cross the road 16 
Pedestrian take long time to cross the road 8 
Expect pedestrian to yield to motorists 3 
4.2.6 Magnitude of Pedestrian Safety Problem 
Respondents were asked if they knew a pedestrian who has been a victim to 
vehicle crash. Forty percent of respondents personally knew a pedestrian who 
has been involved in vehicle-pedestrian crash.  
 
4.2.7 Effectiveness of New Zebra Crossing 
Eighty two percent of 166 respondents think that marking crossings will increase 
pedestrian safety. Logistic regression was performed to assess the predictive 
power of gender, age, and driving experience on driver’s view on effectiveness of 
    
  57 
the crossing. As shown in Table 4.20, when all three factors are included in the 
model (model 1), only the age factor is significant. High correlation between age 
and driving experience, and higher p-values of greater than 0.1 for gender & 
driving experience prompted dropping the two factors resulting into model 2. 
The constant term was also dropped from model 2 because of its higher p-value, 
resulting into model 3. The final model (model 3) is superior in R
2
, percent 
correctly classified, and -2LL (-2 x Log likelihood) criteria. An odd ratio of 1.045 
(1.033< 95% CI <1.057) in the final model indicates that respondents who are one 
year older are 1.045 times more likely to support the crossing. These results 
indicate that the public appreciate the effectiveness of the project in improving 
pedestrian safety through new zebra crossing in the country. 
Table 4.20:  SPSS output on respondent attributes’ significance in predicting 
their support for the new zebra crossing 
Model Factors Estimate S.E. Wald Sig.  
Nagelkerke 
R
2
 
Correctly 
classified 
-2LL 
Age .111 .056 3.964 .046 
Gender .685 .432 2.517 .113* 
Experience -.076 .061 1.540 .215* 
1 
Constant -2.363 1.251 3.567 .059 
0.126 82% 143.58 
Age  .045 .016 7.521 .006 
2 
Constant -.044 .582 .006 .939* 
.078 82.8% 156.73 
3 Age .044 .006 59.552 <.001 0.537 82.8% 156.74 
LL-Log Likelihood. 
* dropped from the model.  
 
4.2.8 Survey Reliability 
Interviewing respondents on their driving behaviour is a less expensive method 
of data collection compared with field observations. However such surveys may 
suffer from self-reporting bias and lack of internal reliability problems. The 
extent of these problems is to some extent influenced by survey and 
questionnaire design. Self-reporting bias can be measured by comparing 
respondents’ stated actions and observed actual actions. Internal reliability 
describes consistency of a respondent in answering more than one questions 
    
  58 
demanding similar response. In this study, drivers’ yielding behaviour was used 
to assess both self-reporting bias and internal reliability. 
 
Self-Reporting Bias 
Observed average drivers’ yielding rate at zebra crossings in the field was lower 
than reported rates in the questionnaire, i.e., 80% (Figure 4.1) vs. 97% to 99% 
(Tables 4.12 & 16) respectively. Such a difference could be attributed by self-
reporting bias in the questionnaire survey. Table 4.21 compares self-reported and 
field observed drivers’ yielding rates. The following trends related to self-
reporting are noted:- 
• Reported “stopping” rates are higher than observed rates; 
• Observed “non-yielding” rate are higher than reported rates; 
• Reported higher “stopping” rates than “slowing” may as well reflect self-
reporting bias since stopping for pedestrians would be considered a better 
practice than just slowing down.  
The magnitude of the bias is higher for the question asking drivers’ requirement 
at zebra crossings than the question asking the meaning of zebra markings – a 
possible effect of wording style.  
 
Comparing reported frequencies in Table 4.18 we see a reversal trend between 
self-reported and reporting of fellow drivers between “always” and “sometimes” 
categories. This reverse of trend may signify self-reporting bias, as the driver 
tends to think that he/she yields for pedestrian more than anyone else. 
Alternatively data in Table 4.18 was analysed by matching respondent’s 
responses resulting in a two-way contingency table (Table 4.22). Highlighted 
diagonal cells of Table 4.22 makes up 39.4% of respondents who gave same 
answers for their behaviour as well as for fellow drivers. Cells above this 
diagonal make up 17.3% percent of respondents who think fellow drivers yield 
to pedestrians more than them. Cells below the diagonal make up 43.2 % of 
    
  59 
respondents who think they yield more frequent than other drivers probably due 
self-reporting bias. Self-reporting bias has also been reported in similar surveys 
such as for seatbelt use (Paradaa et al., 2001; Chartterjee et al., 1991; Robertson, 
1992).  
 
Table 4.21: Comparison between self-reported and observed drivers’ yielding 
behaviour 
Self-reported yielding (%) 
Yielding 
Category 
Meaning of zebra 
markings (Table 4.12) 
Driver requirement at a 
zebra crossing (Table 4.16) 
Observed 
yielding rate (%) 
(Figure 4.1) 
Stopping 83 92 14 
Slowing 14 7 66 
Nothing 3 1 20 
  
Table 4.22: Self-reporting and reporting others yielding behavior  
Other drivers 
   
never sometimes Always 
never 1(0.5%) 3(1.6%) 4(2.2%) 
Sometimes 3(1.6%) 47(25.4%) 25(13.5%) 
Self-
reported 
Always 3(1.6%) 74(40%) 25(13.5%) 
 
Internal Consistency 
A driver that yields for a pedestrian who is waiting to cross the road will likely 
yield to a pedestrian who is in the crosswalk as well. This analogy allows 
reclassification of the five response categories on the “meaning of zebra 
markings” (Table 4.12) to only three categories similar to those used in the 
question on drivers’ requirement at a zebra crossing (Table 4.16). These two 
questions become technically similar because they demand similar answers. The 
difference in wording and phrasing makes difficult for the respondent to detect 
their similarity. Comparison of responses between these two questions allows an 
assessment of respondents’ consistency shown in Table 4.23. Eighty percent of 
respondents were consistent on their responses between the two questions. The 
80% consistency is made up of the sum of top-left to bottom-right shaded 
    
  60 
diagonal cells of Table 4.23. The respondent was considered consistent if he/she 
provided the same response for the two questions. 
 
Table 4.23: Respondent’s consistency between “meaning of zebra markings” & 
“driver’s requirement at zebra crossing”. 
Reported driver requirement for a pedestrian 
waiting to cross at a zebra crossing 
Reported meaning 
of zebra markings 
to a driver Stopping Slowing Nothing 
Stopping 145 (77.5%) 9 (4.8%) 0 (0%) 
Slowing 22 (11.8%) 3 (1.6%) 2 (1.1%) 
Nothing 3 (1.6%) 2 (1.1%) 1 (0.5%) 
 
Reasons of why respondents don’t yield for pedestrian were matched with 
“meaning of zebra markings” as redefined in Table 4.23 and “driver requirement 
at zebra crossings” as follows:  
• Not seeing the pedestrian in advance as the reason for not yielding to 
pedestrians was matched with understanding of zebra markings and driver 
requirement at zebra crossing to be “stopping”; 
• “Slowing is enough for a pedestrians to cross the road” as the reason for not 
yielding to pedestrians was matched with understanding of zebra markings 
and driver requirement at zebra crossing to be “slowing”; 
• “Pedestrians take long time to cross the road” & “expect pedestrian to yield 
to motorists” reasons for not yielding to pedestrians were matched with 
understanding of zebra markings and driver requirement at zebra crossing to 
be “nothing”. 
Table 4.24 shows the matching between meaning of zebra markings & reasons 
for not yielding to pedestrians resulting into 66% respondents’ consistency. Table 
4.25 shows the matching between driver’s requirement at zebra crossing & 
reason for not yielding to pedestrians with 70% consistency. Consistency rates 
obtained in this study (66% to 80%) are comparable with consistency rates 
obtained elsewhere. Williams et al. (2001) obtained a consistency rate of 70%. 
 
    
  61 
Table 4.24: Respondent’s consistency between “meaning of zebra markings” & 
“reason for not yielding to pedestrian at zebra crossing”. 
Meaning of zebra markings 
Reason for not stopping for a pedestrian 
Stop Slow Nothing 
Don’t see the pedestrian in advance 99(61.9%) 16(10%) 2(1.2%) 
Slowing is sufficient for pedestrian to cross 22(13.8%) 5(3.1%) 0(0%) 
“Pedestrians take long time to cross” &  
“Expect pedestrians to yield for motorists” 
11(6.9%) 3(1.9%) 2(1.2%) 
 
Table 4.25: Respondent’s consistency between “driver’s requirement at zebra 
crossing” & “reason for not yielding to pedestrian at zebra 
crossing”. 
Driver requirement at crossing 
when pedestrian waiting to cross Reason for not stopping for a pedestrian 
Stop Slow Nothing 
Don’t see the pedestrian in advance 109(67.7%) 9(5.6%) 0(0%) 
Slowing is sufficient for pedestrian to cross 23(14.3%) 3(1.9%) 0(0%) 
“Pedestrians take long time to cross” & 
“Expect pedestrians to yield for motorists” 
14(8.7%) 2(1.2%) 1(0.6%) 
 
4.3 Field Visits and Pedestrian Survey 
Field visits to the 21 priority-pedestrian crossings and interview with four school 
principals and two school crossing guards were the sources of unpublished 
information about zebra crossings application and usage in Trinidad.   
 
4.3.1 Pedestrian Views 
The Principal of Maraval RC Primary School has the opinion that the priority-
pedestrian crossing at the vicinity of the school was not effective as expected, 
because drivers do not yield to pedestrian on the crossing to Principal’s 
satisfaction. The school has approached the Police Department to provide police 
officers to assist school children when crossing the road. However, the police do 
not have enough personnel to assist the school. As a precautionary measure, the 
principal have asked children to avoid the crossing by waiting to be picked by 
drivers at the school compound. 
 
    
  62 
The crossing guard at the priority crossing located at St. Aidan’s Anglican 
Primary School along Eastern Main Road (EMR) in Arouca believes that he is at 
work because the school administration does not trust motorists at this 
pedestrian crossing. The Guard has observed that women and young drivers of 
Indian descent do not respect pedestrians on the crossing. The assisted crossing 
at this location is during morning school-opening and afternoon school-closing 
hours. 
 
The Principal of St. Pius Boy’s Primary School in Arouca, through casual 
observations, revealed that drivers do not yield to pedestrians at the crossing. 
The crossing at this location was installed after principal’s lobby at the MoW&T. 
It crosses high traffic volume Lapinot road, an arterial road for accessing Lapinot 
district. The road is the link between EMR and the Priority Bus Route (PBR) two 
of the three major arterials in the highly congested East-West corridor. Lapinot 
road is also used by high speed prison vehicles to/and from Lapinot district. 
Traffic level and characteristics on Lapinot road makes it a dangerous road to 
cross on foot. The principal is planning to request a crossing guard and road 
humps at the location in the event the pedestrian accident happens. 
 
The Principal of Boissierre Village RC Primary School observed that, drivers do 
not yield to pedestrians on the crossing that is located in front of the school. He 
plans one day to dart, at the crossing and see what the drivers will do. The 
Principal observed that apart from the problem of crossing the road, there are 
other traffic problems at the site, such as disrespect of one-way regulation in the 
morning up to 8:30 AM. The principal has witnessed several serious accidents at 
this section, although not involving pedestrians. In one incident the vehicle hit 
the bench outside the school where school children would have been sitting, 
luckily there were no students at that time. In another incident, the vehicle hit the 
school fence. 
    
  63 
  
The Principal of Sevilla Private Primary School informed the research team that 
school children don’t use the crossing located near the school entrance. All 
students are dropped and picked-up at the school compound. He knows many 
non-pedestrian traffic crashes at the area, and therefore he proposes traffic lights 
instead of yellow flashing lights. As discussed in subsection 4.3.2, the Principal 
perceives flashing lights at the crossing as substitute of normal traffic signals 
which regulate traffic between conflicting vehicular traffic streams. 
 
The priority-pedestrian crossing at San Fernando RC Primary School spans a 2-
lane, one-way street in the urban environment. The crossing has no zigzag lines. 
The school security guard has observed that drivers do not stop for pedestrians 
at the crossing. In some cases when a driver stops for a pedestrian, the following 
driver overtakes within the crossing area or a driver in the adjacent lane 
continues without stopping (multiple threats). One school child was hit when 
crossing the road after flashing lights were installed there.  
 
One pedestrian at the Barataria site commented that drivers were not stopping 
for pedestrians at that crossing. On the use of yellow or red light colours at 
signalized intersection, he was of the view that flashing lights should be of red 
colour instead of yellow.  
 
4.3.2 The Crossing at Sevilla Private Primary School 
The crossing is located within the vicinity of a complex five-leg intersection in a 
rural set-up. The crossing crosses the road with high-speed, high-traffic volume 
with significant proportion of heavy trucks to- and from Point Lisas Industrial 
Estates. Short gaps in the main traffic and lack of turn bays on the main road 
approaches results in many conflicts between turning traffic and through traffic. 
There are no footpaths for pedestrian, nor did we see any pedestrian during the 
period we visited the crossing. The land in the vicinity of the crossing is so 
    
  64 
underdeveloped to generate significant pedestrian traffic. No surprise, some 
people like the Principal of Sevilla Private Primary School do not relate this 
crossing with pedestrian activities. 
 
4.3.3 Location of Priority-Pedestrian Crossing 
Field visits revealed how decision making on where to place a priority-
pedestrian zebra crossing could be challenging. Visibility of the crossing in terms 
of adequate driver’s sight distance, short distance to cross the roadway, and 
pedestrian desire lines have to be taken into consideration. Visibility and short 
distance addresses the issue of crossing safety, while pedestrian desire lines, 
which depends on local land use and public transport vehicle stops addresses the 
issue of where the crossing will be highly utilised. Some agencies use roadway 
and traffic characteristics when developing guidelines for crossing location. The 
MoW&T requires that these crossings be located on low speed roads (below 65 
kph). This being the only documented criterion, much of the crossing location 
decision depends on engineering judgement of the Traffic Engineer. 
 
4.3.4 Installation of Priority-Pedestrian Crossing 
Most crossings contain all elements, i.e., flashing lights mounted on strip painted 
posts, pavement markings (zebra and zigzag), and advance signs. However, in 
some cases one or two elements were missing or inappropriately installed such 
as:- 
• Lacking pavement markings (Figure 4.4); 
• Lacking flashing lights (Figure 4.5); 
• One post not strip painted black & white (Figure 4.6); 
• Lacking of advance sign; 
• Zigzag lines painted over solid straight lines (Figure 4.7). 
    
  65 
 
Figure 4.4: Priority crossing without pavement markings 
 
Figure 4.5: Priority crossing without flashing lights 
    
  66 
 
Figure 4.6: Non-striped post(s)  
Figure 4.7: Straight line painted over zigzag line 
    
  67 
4.3.5 Motorists’ Behaviours 
In this section we discuss some observed motorists’ behaviours other than 
drivers’ yielding behaviour. The most common misuse at the crossing was 
disrespect of zigzag lines. In many instances vehicles were seen parking, 
stopping and overtaking within zigzag lines at the crossings (Figures 4.8, 4.9, 
4.10), despite restriction of parking at pedestrian crossings by article 54 of the 
Highway Code (Ministry of Works, 1972). Tall commercial vehicles parked at a 
crossing not only hinder pedestrians in the crossing, they could also hinder the 
visibility of flashing lights.  
 
 
  
 
Figure 4.9: Stopping within the crossing 
Figure 4.8 Parking within the crossing 
    
  68 
Figure 4.10: Overtaking within the crossing 
    
  69 
 
CHAPTER FIVE: CONCLUSION & RECOMMENDATIONS 
The following are the main conclusions and recommendations based on the 
results of this study: 
 
5.1 Conclusions 
1.   Flashing light operation was associated with increased drivers’ yielding to 
pedestrian behaviour irrespective of perceived incipient conflict with 
pedestrian, vehicle size, and vehicle registration use. However, some 
stakeholders, especially school principals think the improvement is not 
satisfactory. 
2.   Flashing light operation appears to have higher effect for stopping condition 
than for yielding (at least slowing down). 
3.  Flashing lights have attained successful levels of drivers yielding to 
pedestrians on the crosswalks; however there is a room for improvement 
because higher yield rates have been attained elsewhere.  
4. Pedestrian walking direction had a higher influence on drivers’ yielding 
behaviour, followed by flashing light operation. Vehicle use and size had no 
significant effect on drivers’ yielding behaviour.  
5. Priority-pedestrian zebra crossing is likely to be more beneficial where 
driving culture marginalize pedestrians and where level of road maintenance 
is inadequate. 
6.  The safest and most effective pedestrian crossings often use several traffic 
control devices or design elements to meet the information and control needs 
of both motorists and pedestrians (CUTC, 2000). Likewise priority-pedestrian 
crossing to a great extent solves most of visibility problems facing 
conventional zebra crossing because it uses several traffic control devices. 
7.  The public appreciate the effectiveness of marking pedestrian crossings to 
alleviate pedestrian safety problem, more so by older drivers. 
    
  70 
8. Zigzag lines at priority-pedestrian zebra crossing are highly misunderstood 
by drivers, this is associated with frequent vehicles parking, stopping, and 
overtaking within the crossing area thereby jeopardizing pedestrian safety. 
This case demonstrates a phenomenon in which engineers would design and 
place traffic control devices, which are not properly understood by road 
users. 
9. Drivers’ reported yielding behaviour from this study, although reliable 
should be reviewed in light of possible non-response bias. Non-response bias 
is a systematic tendency for selected elementary units with particular 
characteristics not contributing information while other such units, with other 
characteristics do.   
 
5.2 Recommendations 
1. Drivers’ misunderstanding of zigzag lines zebra markings, and flashing lights 
signifies the need for more education and public information on the crossing 
features and its use. Furlonge (2004) suggested that: “before all the fancy zigzag 
traffic markings and wigwag traffic signals are installed, there must be a strategic 
traffic safety education programme, identifying the purpose of these measures; 
otherwise there is simply a waste of materials and the risk of endangering users of the 
crosswalks”.   
2.  Observed parking violation at priority pedestrian crossings may need 
mitigations more than public information and education. Supplementing 
zigzag lines with standard signs which are familiar to most drivers in T&T for 
restricting parking and stopping; addressing the demand of parking in the 
vicinity of a crossing are among considerations. 
3. The length of “no-parking” zone at priority pedestrian crossing need to be 
reviewed, the current length of up to eight metres appear too short compared 
with other agencies worldwide. 
    
  71 
4. There is a need to develop comprehensive guidelines for selection of 
treatment type at specific crossing location, taking into consideration traffic 
level, number of lanes, and roadway speed limit. 
5. There is a need to formulate and implement enforcement at priority-
pedestrian crossings. 
6. For the crossing to command attention from drivers, installation should 
include all traffic controls in the design specifications, i.e., flashing lights, 
pavement markings, advance signs, and reflectorization of flashing light 
posts. 
7. There is a need to differentiate flashing lights at signalised intersection and 
those at priority crossings in order to create and preserve a unique image for 
this crossing. Activated flashing lights with pedestrian pictorial inscription at 
intersection were found to increase driver yielding behaviour (Van Houten et 
al., 1998). 
8. The crossing could further be improved by providing pedestrian activation 
system, use of red colour instead of yellow, use of advance stop lines, use 
advance signing, and use of LED instead of incandescent because of its 
brighter and economical in long term. Advance signing instructing drivers to 
yield to pedestrian when flashing lights are activated were found to increase 
driver yielding behaviour at intersections (Van Houten et al., 1998). 
9. Overhead flashing lights could solve visibility problem that might be created 
by tall trucks parking in the vicinity of the crossing or overgrown tree 
branches.  
10. Use of solar energy instead of standard power would free the crossing from 
power cuts and also make them feasible in some rural areas where there is no 
conventional power supply. 
11. Study on pedestrian behaviour at priority pedestrian crossings is 
recommended in order to qualify their effects on pedestrians. 
    
  72 
12. There is a need to repeat field observations on drivers’ yielding behaviour to 
determine the long-term effect of this crossing. 
13. Because of positive effect by flashing lights on drivers’ yielding behaviour 
and positive opinion by drivers on the effectiveness of the new crossing, the 
government should continue her current effort to install the new pedestrian 
crossing. 
    
  73 
REFERENCES 
Australian Road Research Board (1989). Understanding Traffic Control Devices, 
ARRB, Special Report No. 44.  
Campbell, B.J., C.V. Zegeer, H.H. Huang, and M.J. Cynecki. (2004). A Review of 
Pedestrian Safety Research in the United States and Abroad. Federal 
Highway Administration. Final Report No. FHWA-RD-03-042. January 2004. 
(http://www.walkinginfo.org/pdf/PedSynth/Ped_Synthesis_Report.pdf) 
Center for Urban Transportation Research (CUTR) (2000). Making Crosswalks 
Safer for Pedestrians: Application of a Multidisciplinary Approach to 
Improve Pedestrian Safety at Crosswalks in St. Petersburg, Florida, 
University of South Florida, Tampa, Florida, July 2000. 
Chartterjee, A., G. Evans, S.H. Richards, and H. Hafford. (1991). Safety Belt Use 
in Tennessee. In ITE Journal, Vol. 61, No.1, pp. 23-27. 
Chu, X., M. Guttenplan, and M. Baltes. “Why People Cross Where They Do.” 
TRB Paper No. 03-3078, In CD-ROM of 82
nd
 TRB annual Meeting, 
Transportation Research Board. Washington, DC, September 2002. 
Clark, K. L., J. E. Hummer, and N. Dutt (1996). Field Evaluation of Fluorescent 
Strong Yellow-Green Pedestrian Warning Signs. In Transportation Research 
Record: Journal of the Transportation Research Board, No. 1538, TRB, National 
Research Council, Washington, D.C., pp. 99-46. 
Connely, M. L., H. M. Conaglen, B.S. Parksonson, and R.B. Isler (1998). Child 
Pedestrians’ Crossing Gap Thresholds. In Accident Analysis & Prevention, Vol. 
30, No.4, pp. 443-453. 
Cooper, P. J., M. Pinili, and W. Chen. (1995). An Examination of the Crash 
Involvement Rates of Novice Drivers Aged 16 to 55. In Accident Analysis & 
Prevention, Vol. 27, pp. 89-104. 
Davies, D.G. (1999). Research, Development, and Implementation of Pedestrian 
Safety Facilities in the United Kingdom. Federal Highway Administration. 
Report No. FHWA-RD-99-089. December 1999. 
    
  74 
Defour, A. (2004). Pedestrians Have Right of Way at Zebra Crossings. Express 
Newspaper, May. 6
th
, 2004.  
Department of Transport (2007). UK Highway Code. 
http://www.direct.gov.uk/en/TravelAndTransport/Highwaycode/DG_070
339 
DeVeauuse, N., K. Kim, C. Peek-Asa, D.L. McArthur, and J.F. Kraus. (1999). 
Driver compliance with stop signs at pedestrian crosswalks on a University 
campus. Journal of American College Health Vol. 47, pp. 269-274. 
Dunn, R.C.M. (1989). Un-signalized Pedestrian Crossings: New Zealand 
Technical Recommendations. In ITE Journal, Vol. 59, No.9, pp. 19-23.  
Ekman, L., and C. Hyden (1999). Pedestrian Safety in Sweden. Federal Highway 
Administration. Report No. FHWA-RD-99-091. December 1999. 
Express
a 
(2004). No Need for Highway Guard. Express Newspaper, May 6
th
 2004. 
Express
b
 (2005). Traffic Light Threat for Debe High School. Express Newspaper, 
February 21
st
 2005.  
Federal Highway Administration (1988). Manual of Uniform Traffic Control Devices 
for Streets and Highways. Part I. FHWA-SA-89-006, U.S. Department of 
Transportation, Washington, D.C.  
Fitzpatrick, K., S. Turner, M. Brewer, P. Carlson, B. Ullman, N. Trout, E. S. Park, 
J. Whitacre, N. Lalani, and D. Lord (2006). Improving Pedestrian Safety at Un-
signalized Crossings: Chapter 3-- Review of Pedestrian Crossing Treatments. 
TCRP Report 112/NCHRP Report 562. Transportation Research Board. 
Washington, D.C. 
Fitzpatrick, K., S. Turner, M. A. Brewer (2007). Improving Pedestrian Safety at 
Unsignalized Intersections. In ITE Journal, Vol. (May 2007), No.7, pp. 34-41. 
Ford, G.L., and D.L. Picha (2000). Teenage Drivers’ Understanding of Traffic 
Control Devices. In Transportation Research Record, 1708, TRB, National 
Research Council, Washington, D.C., pp. 1-11. 
Furlonge, R.  (2004). Protecting Pedestrians. Express Newspaper, March 25
th
 2004. 
    
  75 
Government of Trinidad & Tobago
a
 (1993). Legal Notice No. 1, Legal Supplement 
Part B, Vol. 32, No.1. 4
th
 January 1993. 
Government of Trinidad & Tobago
b
 (2001). Legal Notice No. 241, Legal Supplement 
Part B, Vol. 40, No.209. 31
st
 October 2001. 
Guardian (2004). Pedestrian—Priority Zebra Crossings. Guardian Newspaper, 
April. 30
th
, 2004. 
Harré, N. (2003). Discrepancy  Between Actual and Estimated Speeds of Drivers 
in the Presence of Child Pedestrians. In Injury Prevention.Vol.9, pp. 38–41. 
Hemenway, D. and S.J. Solnick (1993). Fuzzy Dice, Dream Cars, and Indecent 
Gestures: Correlates of Driver Behaviour? In Accident Analysis & Prevention, 
Vol. 25, pp. 161-170. 
Inman, V.W., and G.W. Davis (2007). Synthesis of Literature Relevant to 
Roundabout Signalization to Provide Pedestrian Access. Final Report. 
http://www.access-board.gov/research/roundabouts-signals/report.pdf. 
Jones, T.L., and P. Tomcheck (2000). Pedestrian Accidents in Marked and 
Unmarked Crosswalks: A Quantitative Study. In ITE Journal, Vol. 70, No.9, 
pp. 42-46.  
Knutton, M. (2004). The Best Level Crossing is one that Doesn't Exist: A Month in 
the Life of a Finnish Level Crossing was Captured on a "Peeping Tom" Video 
to Record the Number of Violations by Motorists, Cyclists, and Pedestrians. 
In International Railway Journal, April 2004.  
Koepsell, T., L. McCloskey, M. Wolf, A.V. Moudon, D. Buchner, J. Kraus, and M. 
Patterson (2002). Crosswalk Markings and the Risk of Pedestrian-Motor 
Vehicle Collision in Older Pedestrians. In Journal of American Medical 
Association. Vol. 288, No. 17. Pp. 2136-2143. 
Land Transport NZ (2006). Pedestrian Crossings ? Requirements. Traffic Note 1 
– Revision 2. September 2006. 
    
  76 
Leden, L., P. Garder, and C. Johansson (2006). Safe Pedestrian Crossings for 
Children and Elderly. In Accident Analysis & Prevention, Vol. 38, No.2, March 
2006. pp. 289-294. 
Lee, J., T.L. Maleck, and W.C. Taylor (1999). Pavement Marking Material 
Evaluation Study in Michigan. In ITE Journal, Vol. 69, No.7, pp. 44-51.  
Lesch , M.F. and P.A. Hancock (2004). Driving Performance During Concurrent 
Cell-Phone Use: Are Drivers Aware of Their Performance Decrements? In 
Accident Analysis & Prevention, Vol. 36, No.3, May 2004. pp. 471-480. 
Litman, T.,  R. Blair, B. Demopoulos, N. Eddy, A. Fritzel, D. Laidlaw, H. Maddox, 
and K. Forster (2002). Pedestrian and Bicycle Planning: A Guide to Best 
Practices. 11 October, 2002. (http://www.vtpi.org/nmtguide.doc). 
MacGregor, C., A. Smiley, and A. Dunk (1999). Identifying Gaps in Child 
Pedestrian Safety: Comparing what Children do with what Parents Teach. In 
Transportation Research Record, 1674, TRB, National Research Council, 
Washington, D.C., pp 32-40. 
Metropolitan Transportation Commission (MTC) (2007). European Pedestrian 
and Bicycle Crossing Improvements. 
http://www.mtc.ca.gov/planning/bicyclespedestrians/tools/european/ind
ex.htm 
Ministry of Works (1972). Trinidad & Tobago Highway Code. Port of Spain, 
Trinidad. 
Mutabazi, M., and A. Ackberali. (2008). Education at All Costs: A Case of 
Trinidad School Transport. In CD-ROM of 87
th
 TRB Annual Meeting. 
Transportation Research Board. Washington, DC. 
Nitzburg, M. and R. L. Knoblauch (2001). An Evaluation of High-Visibility 
Crosswalk Treatments—Clearwater, Florida. Research Report No. FHWA-
RD-00-105. Federal Highway Administration. August 2001. 
Paediatrics for Parents (1991). Crossing the Street - Pedestrian Deaths Among 
Children”. April 1991. 
    
  77 
Paradaa, M.A., L.D. Cohn, E. Gonzaleza, T. Byrd, and M.Cortes (2001). The 
Validity of Self-Reported Seatbelt Use: Hispanic and Non-Hispanic Drivers in 
El Paso. In Accident Analysis & Prevention. Vol. 33, No.1, pp. 139-143. 
Peters, K. (2004). Pedestrians Have Limited Right of Way”. Guardian Newspaper, 
May 11
th
 2004.  
Public Works (2003). Signs Alleviate Long-Term Crosswalk Safety Problem. In 
Public Works Journal. Vol. 134, No.2, February 2003, pp.34-36. 
Queensland (2002). Pavement Markings on Approach to Pedestrian Crossings. Issue: 
2/9/2002.  
Räsänen, M., T. Lajunen, F. Alticafarbay, and C. Aydin (2007). Pedestrian Self-
Reports of Factors Influencing the Use of Pedestrian Bridges. In Accident 
Analysis & Prevention, Vol. 39, No. 5, pp. 969-973. 
Redmon, T. (2003). Assessing the Attitudes and Behaviours of Pedestrians and 
Drivers in Traffic Situations. In ITE Journal Vol. 73. No.4, pp.26-30. 
Retting, R.A., A. S. Ferguson, and A.T. McCartt (2003). A Review of Evidence-
Based Traffic Engineering Measures Designed to Reduce Pedestrian–Motor 
Vehicle Crashes. In American Journal of Public Health. Vol. 93, No.9, pp. 1456–
1463.  
Ribbens, H. (1996). Pedestrian Facilities in South Africa: Research and Practice. In 
Transportation Research Record, 1538, TRB, National Research Council, 
Washington, D.C., pp 10-18. 
Robertson, L.S. (1992). The Validity of Self-Reporting Behavioural Risk Factors: 
Seatbelt and Alcohol Use. In The Journal of Trauma. Vol. 32, No.1, 1992, pp. 58-
59. 
Ruden, R. J., and J. Coleman (1979). Grade Crossings Active Warning Devices. In  
ITE Journal Vol. 49, No.3., pp. 23-25. 
Russell, E.R., and W. Kent. (1993). Highway-Rail Crossing Safety 
Demonstrations: Final Report. Federal Railroad Administration, U.S. 
Department of Transportation. 
    
  78 
Rys, M., E. R. Russell, and A. Rys (2001). Characteristics and Luminance Needs 
from Overhead Guide Sign Illumination from Vehicular Headlamps. 2001 
Annual Conference & Exhibition of the Transportation Association of 
Canada; Halifax, Nova Scotia; Canada; 16-19 Sept. 2001. pp. 1-22. 
 Shinar, D; and R. Compton (2004). Aggressive Driving: An Observational Study 
of Driver, Vehicle, and Situational Variables. In Accident Analysis & Prevention, 
Vol. 36, No. 3, May 2004. pp. 429-437. 
Shurbutt, J., R. Van Houten, and T. Shawn (2008). An Analysis of the Effects of 
Stutter Flash LED Beacons to Increase Yielding to Pedestrians Using 
Multilane Crosswalks. In CD-Rom of 87
th
 TRB annual meeting. 
Transportation Research Board. Washington, D.C. 
Smadi, O., R.R. Souleyrette, D.J. Ormand, and N. Hawkins (2008). An Analysis of 
the Safety Effectiveness of Pavement Marking Retroreflectivity. In CD-Rom of 
87
th
 TRB annual meeting. Transportation Research Board. Washington, D.C. 
Smith, K. L. (2008).  Driver-Focused Design of Retroreflective Sheeting for Traffic 
Signs. In CD-Rom of 87
th
 TRB annual meeting. Transportation Research 
Board. Washington, D.C. 
Sparks, J. W., and M.J. Cynecki (1990).Pedestrian Warning Flashers in an Urban 
Environment do they Help? . In ITE Journal Vol. 60, No.1., pp. 32-36. 
St. Bernard, G. and W. Matthews (2003). A Contemporary Analysis of Road 
Traffic Crashes, Fatalities, and Injuries in Trinidad and Tobago. In Injury 
Control and Safety Promotion. Vol.10, No.1-2, pp 21-27. 
Stokes, R.W., M.J. Rys, and E.R. Russell (1996).  Motorists Understanding of 
Selected Warning Signs. In ITE Journal, Vol.66, No.8, 1996. pp.36-41. 
Tan, C. H., and C.V. Zegeer (1995). European Practices and Innovations for 
Pedestrian Crossings. In ITE Journal Vol. 65, No. 11., pp. 24-31. 
Thompson, S.J., E.J. Fraser, and C.I. Howarth (1985). Driver Behaviour in the 
Presence of Child and Adult Pedestrians. In Ergonomics, Vol. 28, No. 10. pp. 
1469-1474. 
    
  79 
Turner, S., K. Fitzpatrick, M. Brewer, and E.S. Park (2006). Motorist Yielding to 
Pedestrians at Unsignalized Intersections: Findings form a National Study on 
Improving Pedestrian Safety. In CD-ROM of 85
th
 TRB Annual Meeting. 
Transportation Research Board. Washington, D.C. 
Ullman, B., K. Fitzpatrick, and N. Trout (2004). On-Street Pedestrian Surveys of 
Pedestrian Crossing Treatments. Institute of Transportation Engineers’ 
Annual Meeting and Exhibit. Lake Buena Vista, Florida. 
Underwood, J.R. and T.M.A. Rasul, (1984). A Survey of Travel by Children to 
School in the Curepe/St. Joseph Area of Trinidad. In WIJE, Vol. 9 No.2, pp. 
73-78. 
Van Houten, R., K. Healey J. E. L. Malenfant, and R. Retting (1998). Use of Signs 
and Symbols to Increase the Efficacy of Pedestrian-Activated Flashing 
Beacons at Crosswalks. In Transportation Research Record, 1636, TRB, National 
Research Council, Washington, D.C., pp 92-95. 
Van Houten, R., J. E. L. Malenfant, and D. McCusker (2001). Advance Yield 
Markings: Reducing Motor Vehicle–Pedestrian Conflicts at Multilane 
Crosswalks with Uncontrolled Approach. In Transportation Research Record, 
1773, TRB, National Research Council, Washington, D.C., pp 69-74. 
Van Houten, R., D. McCusker, S. Huybers J. E. L. Malenfant, and D. Rice-Smith 
(2002). Advance Yield Markings and Fluorescent Yellow-Green RA 4 Signs at 
Crosswalks with Uncontrolled Approaches In Transportation Research Record, 
1818, TRB, National Research Council, Washington, D.C., pp 119-124. 
Van Houten, R., and J. E. L. Malenfant (2001).  ITS Animated LED Signal Alert 
Drivers to Pedestrian Threats. In ITE Journal Vol. 71, No.7, pp. 42-47. 
Van Winkle, J.W., and D.A. Neal. (2000). Pedestrian-Actuated Crosswalk 
Flashers, Paper presented at the ITE 2000 Annual Meeting, Institute of 
Transportation Engineers, Nashville, TN. August 2000. 
    
  80 
Várhelyi, A., (1996). Dynamic Speed Adaptation Based on Information 
Technology — a Theoretical Background. Bulletin 142, Department of Traffic 
Planning and Engineering, Lund Institute of Technology. Sweden. 
Várhelyi, A. (1998). Drivers’ Speed Behaviour at a Zebra Crossing: A Case Study. 
In Accident Analysis & Prevention, Vol. 30, No. 6, pp. 731-743. 
Vermont Agency of Transportation (2004). Guideline for the Installation of 
Crosswalk Markings and Pedestrian Signing. Revised 2004. 
Williams, L., T. Sobieszczyk, and A.E. Perez (2001). Consistency between Survey 
and Interview Data Concerning Pregnancy Wantedness in the Philippines. In 
Studies in Familly Planning Journal, Vol. 32, No.3, pp. 244-253. 
Yeh, M. and J. Multer (2007). Traffic Control Devices and Barrier Systems at 
Grade Crossings: A Literature Review. In CD-ROM of  the 86
th
 TRB Annual 
Meeting. Transportation Research Board, Washington, D.C. 
Zegeer, C.V., C.T. Esse, J.R. Stewart, H.F. Huang, and P. Lagerwey. (2004). Safety 
Analysis of Marked Versus Unmarked Crosswalks in 30 Cities. In ITE Journal, 
Vol.74, No.1, pp. 34-41. 
    
  81 
APPENDIX: QUESTIONNAIRE