Open Journal of Civil Engineering, 2013, 3, 39-45
http://dx.doi.org/10.4236/ojce.2013.33B007 Published Online September 2013 (http://www.scirp.org/journal/ojce)
Copyright © 2013 SciRes. OJCE
Pedestrian Countdown Signals: What Impact
on Safe Crossing?
Janusz Supernak, Vinay Verma, Iga Supernak
Department of Civil, Construction and Environmental Engineering, San Diego State University, San Diego, USA
Email: supernak@mail.sdsu.edu
Received July 2013
ABSTRACT
This paper examines safety impacts of a Pedestrian Countdown Signal (PCS) installed on a busy downtown intersection
in San Diego, California. Crossing episodes of over 5000 pedestrians were videotaped and analyzed using multivariate
statistical methods. Details of timing of pedestrian crossing as well as information about vehicular traffic and signal
timing were carefully coded for each pedestrian. Significant safety benefits of the PCS system were found on the long
crossings over a street with high vehicular volumes: most pedestrians were able to effectively increase their walking
speed to complete their crossing without committing the exit violation even if they have already committed the entry
violation. However, on the short crossing with light vehicular traffic, PCS was generally ineffective in preventing the
entry violations from becoming exit violations. Over there, many pedestrians felt safe enough to walk over a short
crossing with no apparent vehicular traffic in sight instead of waiting for a green signal. The length of crossing and vo-
lume of interfering vehicu lar traffic were consistently found the most significant variables affecting the crossing viola-
tion rates of different categories of pedestrians. Crossing violation rates were the highest for runners, bicyclists and old-
er males. Crossing violation charac teristics were found to be consistent over time.
Keywords: Pedestrian Safety; Countdown Signals; Multivariate Analysis; Safety Impacts
1. Introduction
PCS systems aim at improving pedestrian safety at the
intersections: information about the number of seconds
remaining for a safe crossing should help pedestrians to
make a correct decision whether to start crossing or to
wait for the next Steady Walk (SW) signal. It could also
alert them to speed up their walking to safely complete
their crossing maneuver in time and avoid any conflict
with the vehicular traffic.
Generally, PCS systems proved to bring tangible bene-
fits in terms of reducing the percentage of finishing of
the pedestrian crossings on Don’t Walk (DW) signal (il-
legal exit). However, several studies indicated that PCS
may actually increase the percentage of pedestrians who
start crossing on Flashing Don’t Walk (FDW) signal (il-
legal entry) as they may realize that they still have
enough time to complete their crossing safely. A large
San Francisco, CA study [1] found that PCS brought an
impressive 52% reduction in pedestrian injury as well as
statistically significant decrease in exit violations and in
running or aborting crossing. At the same time, there was
a slight (and statistically not significant) increase in the
entry violations. The observed vehicle-pedestrian con-
flicts decreased but not to the statistically significant le v-
el. P edestrians were found to increase their speed to clear
intersection in time. Similar positive results were re-
ported in the Monterey, CA [2] and Las Vegas, NV [3]
studies. But in the Montgomery County, study [4], the
PCS performance was found to be somewhat mixed: only
three out of five intersections studied found decrease in
the exit violations. Even less impressive are results from
the Lake Buena Vista, FL study: PCS installation re-
sulted in more rather than less exit and entry violations
there [5]. A before-after study performed at the city-first
PCS-equipped intersection in downtown San Diego
found that the new system brought a statistically signifi-
cant decrease in percentage of illegal exits but also in-
crease (although not statistically significant) in illegal
entries [6].
International studies generally confirm the benefits of
the PCS systems. For example, in Shanghai, China, PCS
installation dramatically increased the proportion of suc-
cessful crossing; the resulting safety improvement was
experienced primarily by the older people [7]. A study
from Auckland, New Zealand, concluded that PCS sys-
tems favorably affect pedestrian crossing behavior if
placed at suitable locations but may lead to increase in
risky crossing behavior at some unsuitable locations [8].
J. SUPERNAK ET AL.
Copyright © 2013 SciRes. OJCE
40
Most commonly, PCS installations begin the count-
down procedure on the start of FDW interval rather than
on the SW interval. However, some cities follow the op-
posite procedure. Although there may be still discussion
on which version of the PCS system may be preferred by
the pedestrians, it is quite clear by now that the pede-
strians overwhelmingly favor intersections equipped with
PCS systems over those without PCS.
Review of the literature indicates some diversity of
conclusions reflecting not only any potential differences
in pedestrian behavior over the wide range of local cir-
cumstances but also some methodological issues related
to study design and the actual analytical methods applied.
This paper reports on an effort undertaken at San Diego
State University (SDSU) to better understand the pede-
strian crossing behavior at an intersection equipped with
a PCS system [9]. Availability of the newly installed
PCS system in downtown San Diego created not only an
incentive to study its effectiveness but also created an
opportunity to look into the core “anatomy” of the pede-
strian crossing behavior through careful recording and
coding of over 5000 crossing episodes made by a diver-
sified gr oup of pedes trians.
2. Study Objectives
The main goal of the SDSU study reported in this paper
was to evaluate the long term effect of pedestrian count-
down signals on the intersection crossing behavior of
various groups of pedestrians in the context of pedestrian
safety. Data collection was conducted using a video
camera. The relevant data from each individual crossing
episode were extracted from the video tapes and coded
into a spreadsheet format suitable for analysis. Finally,
the data were analyzed using statistical methods to sys-
tematically study the pedestrian behavior at that PCS-
equipped intersection.
The following research questions were posted: 1) How
common are pedestrian crossing violations? 2) Who are
the violators? 3) Where do the violations happen? 4)
When do the violations happen? 5) Are violations af-
fected by the geometry of the crossing? 6) Are vio lations
affected by the vehicular traffic? 7) Are violations con-
sistent over time? 8) Do PCS systems reduce the most
dangerous violation: illegal exit? 9) Do pedestrians use
PCS displays to adjust their walking behavior to com-
plete their intersection crossing in time?
3. The Venue
The selected intersection is located at Broadway and
Second Avenue in the downtown San Diego. It is sur-
rounded by Westgate Hotel building, San Diego City
Hall, NBC San Diego, Speckles Theater, and several
stores and restaurants. A major trolley station is nearby.
The study intersection has high pedestrian traffic vo-
lumes during the day. The vehicular traffic volumes are
high on Broadway but light on the Second Avenue. The
study intersection is not equipped with pedestrian push
buttons. The pedestrian crossing cycles are activated
once per traffic cycle. Data was collected in the morning
peak period time 8:00 to 10:00. One crosswalk was se-
lected for one specific day. Video recording was done for
three weeks from Monday to Thursday. Pedestrian video-
tape records were examined carefully to properly code
the behavior of each pedestrian on Excels spreadsheets.
To assure privacy, no specific person was ever targeted
or identified.
Broadway Street is almost two times wider than the
Second Avenue. The Broadway crossing distance is 25.8
m (86 feet) whereas the Second Avenue crossing dis-
tance is 14.4 m (48 feet). At the time of the videotaping
in August 2009, the c ycle length on the study inter se ction
was c = 70 seconds. Cycle phasing for pedestrians was
set as follows:
Broadway Crossing:
Walk: 7 sec; FDW: 21 sec and DW: 42 sec; Total: 70
sec.
Second Avenue Crossing:
Walk: 20 sec; FDW: 10 sec and DW: 40 sec; To tal: 70
sec.
Assuming the conservative walking speed standard of
1.0 m/sec (3.5 ft/sec), a pedestrian would need almost 26
seconds to cross Broadway; a person who starts crossing
at the beginning of the SW indication will have just 28
seconds to complete his/her crossing. If the 1.2 m/sec
(4.0 ft/sec) speed standard was applied, the pedestrian
would need 21.5 second s to cros s. This mean s that a slow
walking pedestrian who enters legally at the end of his/
her SW phase may have difficulty to exit on time since
the FDW signal is just 21 seconds long.
The pedestrian timing plan for the Second Avenue
crossing appears problematic. Assuming the 1.0 m/sec
walking speed, a pedestrian would need almost 14
seconds to cross Second Avenue but the FDW phase
there was only 10 seconds long. Thus, a pedestrian who
starts crossing legally at the end of the SW indication
may not have enough time to exit legally. Even if 1.2
m/sec standard was applied, the pedestrian would need
12 seconds to cross, more than the duration of the FDW
phase. This problematic timing plan on the Second Ave-
nue design (too long SW phase and too short FDW phase)
may well be one of the contributing factors to Exit viola-
tions on that crossing.
4. Pedestrian Data Organization
4.1. Pedestrian Crossing Situations
Every pedestrian’s intersection crossing behavior can be
J. SUPERNAK ET AL.
Copyright © 2013 SciRes. OJCE
41
classified as one of the four following situation s :
Situation 1: Non Violation; Legal E ntry and Le gal Exit.
Situation 2: Illegal Entry but Legal Exit.
Situation 3: Legal Entry but Illegal Exit.
Situation 4: Illegal Entry and Illegal Exit.
In order for the Entry to be legal, it has to start on
Steady Walk (SW) signal. In order for the Exit to be le-
gal, it has to be completed on Steady Walk (SW) or
Flashing Don’t Walk (FDW) signal.
4.2. Definition of Crossing Violations
In order to effectively record and process the observed
incidents of intersection crossing done by each individual
pedestrian, the raw data was organized on a spreadsheet
to record for each pedestrian the following pieces of in-
formation : 1) date; 2) cycle number; 3) crossing identi-
fication (Broadway or Second Avenue); and 4) person
category number.
The following crossing-related information was ex-
tracted from the video for each pedestrian P:
Tw = duration of the Steady Walk (SW) phase (sec).
Tfdw = duration of the FDW phase (sec).
to = time gap between the start of SW phase and the
moment pedestrian P started his/her crossing (sec).
t1 = time used by pedestrian P to reach the median of
the intersection (sec).
t2 = time used by pedestrian P to traverse the distance
between the intersection median and the intersection end
(sec).
The following condition constitutes the Illegal Entry
by pedestr i an P:
w0
Tt0−<
(1)
The following condition constitutes the Illegal Exit by
pedestrian P:
( )
wfdw0 1 2
TTttt 0+− ++<
(2)
If pedestrian P is increasing his/her speed as his/her
intersection crossing progresses, the following condition
is expected to be met:
21
tt0−<
(3)
In order to examine the effect of PCS on safe crossing,
there is a need to identify those pedestrians who have
already committed Entry Violation (entered on FDW)
and were heading for the Exit Violation if their speed
was not adjusted (increased).
For this particular subset of pedestrians the condition
t0 > Tw needs to meet the following additional condition:
01w fdw
t 2t TT+>+ (4)
If their speed was not adjusted, the original time t1
would be repeated on the second half of the intersection.
However if the speed increase had sufficiently reduced t2,
the Exit Violation could be avoided.
Analysis of the intersection crossing episodes made by
various categories of pedestrians at the study intersection
in downtown San Diego proved that the vast majority of
them crossed the intersection with a speed much faster
than the design speed of 1.0 m/sec (3.5 feet/sec). Thus,
despite the rather tight timing plan, Situation 3: legal
entry yet illegal exit proved to be a very rare case. Dur-
ing the three weeks of videotaping of 5504 crossing epi-
sodes, this situation was encountered only once. For that
reason, Situation 3 was not analyzed as a separate case.
Thus, the study intersection exhibited only two distinct
violation cases, consistently used in the analysis pre-
sented in this paper:
1) Entry Violation: Illegal Entr y bu t still Legal Exit.
2) Exit Violation: Illegal Exit following Illegal Entry.
5. Analysis
5.1. Initial Definition of Pedestrian Ca t egories
Twelve pedestrian categories were arbitrarily defined
based on pedestrians’ age (below 18; 18 - 40; 40 - 65,
above 65), gender (male, female) and their specific status
(handicapped, runners, bicyclists). Their crossing beha-
vior was videotaped and coded. Representations of the
defined categories are quite uneven: for example, the
Younger Males Category amounts for a half of all pede-
strians in the sample. Categorization according to age
and handicap was done consistently, based on the best
possible i nt erpreta t ion of the vi deo data.
5.2. Comparison of Violation Proportions for the
Twelve Pedest rian Categories
A total of 5504 pedestrians were observed and classified
into 12 categories. The three performance outcomes:
Non-Violation; Entry Violation; and Exit Violation pro-
portions were examined for the following variables: pe-
destrian categories; time-of-day periods (peak/off-peak);
days of the week (M, Tu, W, and Th); weeks (1, 2, and 3);
length of crossing (short/long) and corresponding vehi-
cular traffic (light, heavy) as well as corresponding tim-
ing plans for the two crossings (short FDW/long FDW).
The starting analysis was performed for the initial
twelve categories of pedestrians to identify similarities
and differences among them in terms of their intersection
crossing behavior. A pair-wise comparisons of the entry
and exit violation proportions was made to evaluate
whether the performance measures between different
those groups are statistically significant. The null hypo-
thesis tested in all cases is that there is no difference be-
tween groups, with the a lter native hypothesis that there is
a statistically significant difference among them.
01 2
Hp p
=: (5)
J. SUPERNAK ET AL.
Copyright © 2013 SciRes. OJCE
42
12
Hp p
a:
(6)
A two-tailed z test was performed at a confidence level
of 95% with the corresponding critical z value (zα/2) of
1.96. Various groups of pedestrians were stratified ac-
cording to several variables, and pair-wise proportion
comparisons were performed to have preliminary know-
ledge about the significance of those variables in de-
scribing variations in Entry and Exit Violations propor-
tions. The z-test brought the following initial results:
1) With both crossings analyzed jointly, older males
were committing the Entry Violation more frequently
than their younger male counterparts and that difference
was significant. The same was also true for the Exit Vi-
olation: older males were exiting illegally with higher
frequency than their younger male counterparts but this
difference was not significant. It is likely that younger
males were able to adjust their walking speed more ef-
fectively than their older counterp arts.
2) Persons performing sports activities (runners, bi-
cyclists) were committing both Entry and Exit Violation
more frequently than their “regular” pedestrian counter-
parts.
3) For both Entry and Exit Violations, there was a sta-
tistically significant difference in their frequency be-
tween short and long crossing. The length of crossing is
just one of the differences between those two crossings.
There are two more differences:
a) The long pedestrian crossing on Broadway is against
heavy vehicular traffic whereas the short crossing over
the Second Avenue has a very light vehicular traffic; and
b) The FDW phase over Broadway is long whereas
FDW phase over the Second Avenue is short.
4) Both Violation types are more frequent during the
peak period (8 - 9 am) than during the off peak period (9
- 10 am), as expected. Rushing to work may be the rea-
son for more frequent violations during the pea k pe riod.
5.3. Comparison of Violation Proportions for the
Six Pedestrian Categories
Initial statistical analysis: comparison of violation pro-
portions for the twelve pedestrian categories helped to
identify those categories of pedestrians who were the
most common violators. However, some of the categories
were not represented with enough observations despite
the relatively long (three weeks) observation period. For
example, people with handicap amounted to just one case
out of 5504 observations; a much larger observation pe-
riod and a larger sample would have been needed to ob-
tain statistically valid results concerning this group of
pedestrians at the study intersection.
For the purpose of the further analysis, some of the
“old” categories 1 through 12 were merged to create six
“new” categories A through F. A common sense judg-
ment was used to create this new categorization of pede-
strians. Definitions of the new categories and respective
Entry and Exit Violation proportions are presented in
Table 1 and can be summarized as follows:
1) Pedestrians crossing the study intersection in down-
town San Diego were predominantly younger males and
females: they accounted for 88% of the entire sample of
pedestrians.
2) The overall Entry Violation proportions were quite
high (around 30%) for both crossings. The highest Entry
Violation rates were for the runners and bicyclists, and
for older males, particularly on the short crossing.
3) The overall Exit Violation proportions were quite
small (around 6%) on the long crossing but remained
substantial on the short crossing (ar ound 20%).
4) On the long crossing, 78% of the Entry Violations
did not lead into consequent Exit Violations. But on the
short crossing, this number was only 34%.
5) Significant reduction in Violation frequency be-
tween the intersection entry and the exit on the long
crossing may be partially attributed to the PCS system;
more detailed analysis follows to test this hypothesis .
5.4. Chi-Square Analysis
Chi-square analysis was performed on the new categories
to evaluate whether pedestrian behavior represented by
the proportions among the three possible outcomes (No
Violation, Entry Violation, Exit Violation) varied with
different pedestrian categories, weeks, time of the day,
day of the week or length of crossing. The null hypothe-
sis tested in all cases is that there is no difference in pro-
portions between the strata, with the alternative hypothe-
sis that there is a statistically significant difference be-
tween the strata:
Table 1. Crossing violation proportions for six person ca-
tegories.
Pedestrian Long crossing Short crossing
Category Entry Exit Ent ry Exit
Violation Violation Violation Violation
A. Children 30.8% 0.0% 8.5% 5.1%
B. Younger Males 29.5% 7.8% 30.0% 20.4%
C. Younger Females 20.6% 3.3% 30.0% 18.6%
D. Older Males 34.1% 3.9% 42.7% 28.9%
E. Older Females 26.1% 4.3% 34.5% 23.8%
F. Runners, Bicyclists
31.6% 5.3% 61.3% 48.4%
Overall 26.6% 5.8% 31.0% 20.6%
J. SUPERNAK ET AL.
Copyright © 2013 SciRes. OJCE
43
01 2
Hp p=
: (7)
12
Hp p
a:
(8)
Results of chi-square analysis are shown in Table 2.
Results that are significant at the 5% level are labeled S,
and not significant are labeled N. The results of the chi
square analysis presented in Table 2 show that o ut of the
five variables studied: pedestrian categories, weeks, days,
time of day, crossing lengthonly “weeks” variable ap-
pear consistently not significant at the 95% confidence
level.
All other variables show statistical significance for at
least some situations listed in the first column. “Length
of crossing” (and related difference in vehicular traffic
and in signal timing) appears to be a strong discriminat-
ing variable. “Person categories” differences also appear
important except for the case when distinction is made
between the Entry Violation and Exit Variation propor-
tions only. “Time of day” is the third significant variable.
Chi-square analysis shows that there is a statistically
significant difference in crossing behavior related to
those three variables when they are analyzed separately.
The next research question is how significant would
those variables be if analyzed jointly. Analysis of va-
riance (ANOVA) is a suitable statistical technique for
this type of analysis.
5.5. Analysis of Variance (ANOVA)
Analysis of variance (ANOVA) analysis was performed
on the new six categories. This test was performed to
evaluate whether the differences in violation proportions
between different categories are statistically significant
when analyzed together with the following variables: the
type of crossing, weeks, day of the week, and time of the
day. All possible 2-way ANOVA calculations were per-
formed separately for the Entry Violation and Exit Viola-
tion cases.
Results of analysis of variance for the Entry and Exit
Violations are presented in a summary way in Table 3
and Table 4, respectively. Differences in proportions that
are significant at the 95% level of confidence are labeled
Table 2. Summary of chi-square analysis results.
Proportion scenarios 1F 2F 3F 4F 5F
VN/VNE/VXE S S N S S
VN/VNE S S N S S
VN/VXE S S N N S
VNE/VXE N S N N S
NV = non-violation proportion; ENV = entry violation proportion; EXV =
exit violation proportion; F1 = pedestrian categories; F2 = crossing length;
F3 = week s; F4 = days; F5 = peak/off-peak; S = signif ican t at 5% level; N =
not significant at 5% level.
Table 3. Entry violation ANOVA significance summary.
srotcaF 1F 2F 3F 4F 5F
1F N S N N
2F S N N N
3F N N N N
4F N N N N
5F N N N N
F1 = pedestrian cate gories ; F2 = crossing length; F3 = weeks; F4 = days; F5
= peak/off-peak; S = significant at 5% level; N = not significant at 5% level.
Table 4. Exit violation ANOVA significance summary.
srotcaF 1F 2F 3F 4F 5F
1F N S S S
2F S S N N
3F N N N N
4F N N N N
5F N N N N
F1 = pedestrian cate gories ; F2 = crossing length; F3 = weeks; F4 = days; F5
= peak/off-peak; S = significant at 5% level; N = not significant at 5% level.
S and those not significant are labeled N.
The main conclusions from the results summarized in
Tables 3 and 4 are as follows:
Two variables show explanatory power to explain
differences in proportions of the Entry Violation be-
havior (entering illegally). Variable “categories” are
significant when analyzed together with variable
“weeks”; and type of crossing is significant only if
analyzed together with variable “categories”. This
result would imply that Entry Violation happens rela-
tively independently of the other variables analyzed.
Exit Violation behavior is quite different than the
Entry Violation behavior. The “category” variable is
significant when analyzed together with variables:
“weeks”, “peak/off peak” and “days” variables (but
only when recorded on the short crossing). The
“crossing” variable is significant when analyzed to-
gether with the “category” and “weeks” variables;
Three remaining variables: weeks, days, time-of-day
do not show explanatory power if analyzed with other
variables. This result sugges ts that pedestr ian cro ssing
behavior is quite consistent over time.
5.6. Impact of PCS on Walking Speed
Some studies indicated that PCS systems may be able to
make pedestrians to adjust their speed of crossing to
avoid the dangerous Exit Violation. Equation (4) identi-
J. SUPERNAK ET AL.
Copyright © 2013 SciRes. OJCE
44
fies those pedestrians who committed Entry Violation
(entered on FDW) and were heading for the Exit Viola-
tion if their speed was not adjusted (increased).
Out of 5402 pedestrians studied, 1084 of them (252 on
Broadway and 832 on the Second Avenue) were those
who committed the Entry Violation and were on track to
also commit the Exit Violation.
One hundred and fifty two (152) out of the 252 pede-
strians (60.3%) who made a long crossing across Broad-
way (high volume of auto traffic), and started late
enough to potentially commit Exit Violation, had in-
creased their speed by observing PCS display and com-
pleted their crossing in time before the steady DW indi-
cation.
Potential for speed adjustment was also studied for the
pedestrians on short crossing (Second Avenue). But in
that case, only 77 pedestrians out of 832 (9.3%) who
committed Entry Violation were able to effectively adjust
their speed to avoid Exit Violation and finish their cross-
ing on time. Two reasons were likely responsible for that:
1) light vehicular traffic that made the short crossing ep-
isode look safe even if it is illegal, and 2) problematic
signal design that offered a very short crossing time for
pedestrians who entered on the Flashing Don’t Walk or
even late on the Steady Walk phase.
5.7. Impacts of Other Fa ctors
5.7.1. Effect of Vehicular Traffic
Pedestrian behavior is naturally affected by the vehicular
traffic density. High vo lume of veh icular traffic may lead
into dangerous pedestrian-vehicle conflicts, and reason-
able pedestrians would try to avoid them. This may be
one of the reasons why pedestrians were committing Exit
Violation on the short Second Avenue crossing more
frequently than on the long and busy Broadway crossing .
Videotape analysis clearly revealed that auto traffic was
much more intense on Broadway than on the Second
Avenue. Gaps between consecutive cars were short on
Broadway, potentially discouraging pedestrians from
making risky crossing decisions. But on the Second Ave-
nue, auto traffic was very light with much larger gaps
between consecutive automobiles. Cars were counted on
the Second Avenue to determine time gaps between them
during the pedestria n FDW and DW phases, and revealed
that 41% of auto gaps there were longer than 8 seconds.
This means that almost every second cycle had zero cars
interfering with the pedestrians on that short crossing.
5.7.2. Enforcement Issue
During the three weeks of the videotaping, there was no
indication of any enforcement action against those who
jaywalked. Lack of enforcement may have been a con-
tributing factor to the magnitude of crossing violations at
the study intersection.
5.7.3. Pl atooning
Some pedestrian crossing violations may be caused by
platooning. When a large group of pedestrian is trying to
cross, some pedestrians may actually block other pede-
strians, delay them and potentially contribute to their
crossing violations. However, no episode of pedestrian
platoon-related delay was observed on either Broadway
or Second Avenue crossing.
6. Summary of Findings
1) Pedestrian Countdown Signals (PCS) appear effec-
tive in reducing exit violations where there is a long
crossing with a long Flashing Don’t Walk phase count-
down (the Broadway case in this study). In case of short
crossing with a short FDW phase, they appear to be ra-
ther ineffective (the Second Avenue case in this study).
Auto traffic may also effect the violations among pede-
strians: for the higher volume auto traffic on the long
crossing fewer violations were recorded as compared to
lighter traffic on the short crossing.
2) The exit violations proportions are greatly reduced
on the long crossing over a major arterial (Broadway).
This reduction is much smaller on the short crossing over
a street with a low traffic volume (Second Avenue).
3) PCS system is proved to influence speed adjustment
of pedestrians who can effectively avoid the consequen-
tial exit violation; this is primarily observed on the long
crossing where 2/3 of pedestrians heading for the exit
violation are effectively able to avoid it and exit the in-
tersection timely.
4) Pedestrian population is heterogeneous in term of
its intersection crossing behavior, and proper market
segmentation appears useful in understanding differences
in violation frequencies among various groups of pede-
strians. In downtown San Diego, the groups with the
highest crossing violation rates are runners and bicyclists
as well as older males.
5) Pedestrian violations are more frequent during the
rush hours than outside that period but this difference is
not statistically significant. Pedestrian intersection cross-
ing behavior is consistent from day to day and from week
to week.
6) Pedestrian behavior is affected by multiple factors
such as composition of the population, intersection geo-
metry, traffic intensity and signal design. In order to
properly capture those various effects, multivariate ana-
lysis is useful.
7) Analysis based on videotaping and detailed coding
of the relevant elements of the intersection crossing of
every individual is a tedious but effective procedure to
better understand subtleties of pedestrian intersection
crossing behavi or.
8) More studies of the same kind would be needed to
verify validity of those findings at some other sites, par-
J. SUPERNAK ET AL.
Copyright © 2013 SciRes. OJCE
45
ticularly at locations with higher percentage of such
groups like: children and adolescents or elderly and han-
dicapped pedestrians.
REFERENCES
[1] F. Markovitz, S. Sciortino, J. Fleck and B. Yee, “Pedes-
trian Countdown Signals: Experience with an Extensive
Pilot Installation,” ITE Journal, 2006.
[2] J. Leonard, M. Juckes and B. Clement, “Safety & Behav-
ior: Behavioral Evaluation of Pedestrians and Motorists
towards Pedestrian Countdown Signals, Report by Des-
sau-Soprin Inc. for the City of Monterey, California,
1999.
[3] S. S. Pulugurtha and S. S. Nambisan, “Effectiveness of
Pedestrian CountdowAZSn Timers to Enhance Safety in
Las Vegas,” Institute of Transportation Engineers District
6 Annual Meeting, Sacramento, June 2004.
[4] J. L. Botha, A. A. Zaby s hny, J. E. Day , R. L. Northous e , J.
O. Rodriguez and T. L. Nix, “Pedestrian Countdown Sig-
nals Study in the City of San Jose: Final Report to the
CaliforniaTraffic Control Devices Committee,” 2002.
[5] H. Huang and C. Zeeger, “The Effect of Pedestrian Count-
down Signals in Lake Buena Vi sta,” Report for the Florida
Department of Transportation, 2000.
[6] J. Supernak and S. Chavez, “Evaluation of Two Pedes-
trian Countdown Systems at the Crosswalks in the City of
San Diego,” Report for the City of San Diego, 2005.
[7] W. J. Ma, Z. Z. Wu and X. G. Yang, “Empirical Analysis
of Pedestrian Countdown Signals in Shanghai,” TRB 88th
Annual Meeting Compendium of Papers, Washington DC,
11-15 January 2009, 11 p.
[8] A. Hooper, V. Vencatachellum and M. Tse, “Trial of Pe-
destrian Signals Incorporating a Numerical Countdown
Display in Auckland CBD,” Presented at the IPENZ
Transportation Group Conference, Tauranga, 2007.
www.ipenz.org.nz/ipenztg/papers/2007/C4a_HooperVenc
atTse.pdf
[9] V. Verma, “Effectiveness of Countdown Pedestrian Sys-
tems in Downtown San Diego,” M.S. Thesis, San Diego
State University, 2012.