Accident Analysis and Prevention 94 (2016) 73–79

Contents lists available at ScienceDirect

Accident Analysis and Prevention

journal homepage: www.elsevier.com/locate/aap

The influence of pedestrian countdown signals on children’s crossing

behavior at school intersections
Lianning Fu ∗ , Nan Zou
School of Control Science and Engineering, Shandong University, 17923 Jingshi Road, Jinan 250061, China

a r t i c l e i n f o a b s t r a c t

Article history: Previous studies have shown that pedestrian countdown signals had different influences on pedestrian
Received 4 March 2016 crossing behavior. The purpose of this study was to examine the effects of the installation of countdown
Received in revised form 15 April 2016 signals at school intersections on children’s crossing behavior. A comparison analysis was carried out on
Accepted 16 May 2016
the basis of observations at two different school intersections with or without pedestrian countdown
Available online 1 June 2016
signals in the city of Jinan, China. Four types of children’s crossing behavior and child pedestrian-vehicle
conflicts were analyzed in detail. The analysis results showed that using pedestrian countdown timers
Keywords:
during the Red Man phase led to more children’s violation and running behavior. Theses violators created
Child pedestrian
Crossing behavior
more conflicts with vehicles. However, pedestrian countdown signals were effective at helping child
Countdown signals pedestrian to complete crossing before the red light onset, avoid getting caught in the middle of crosswalk.
Comparison No significant difference was found in children who started crossing during Flashing Green Man phase
Road safety between the two types of pedestrian signals. Moreover, analysis results indicated that children who
crossed the road alone had more violation and adventure crossing behavior than those had companions.
Boys were found more likely to run crossing than girls, but there was no significant gender difference
in other crossing behavior. Finally, it’s recommended to remove countdown at the end of the Red Man
phase to improve children’s crossing behavior and reduce the conflicts with vehicles. Meanwhile other
measures are proposed to improve children safety at school intersections.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction
Safe crossing is an important aspect of child pedestrian safety,
because most of children are likely to cross at the school intersec-
tions. To improve children’s crossing safety, various measures are
implemented in school zones, such as install high-visibility cross-
ings, speed humps, pedestrian refuge islands, and so on. Moreover,
considering of children are too young to make correct and safe
judgments when crossing, pedestrian signals are usually neces-
sary at the school intersections. The main purpose of pedestrian
countdown signals is to increase pedestrian safety, they are gen-
eral considered to install at signalized pedestrian crossing in school
zones and busy intersections (Lipovac et al., 2013; Lambrianidou
et al., 2013). In recent years, more and more countries are starting
trial to install pedestrian countdown signals at the intersections
to improve pedestrian safety. Generally, pedestrian countdown
signals are used to display the number of remaining seconds at
∗ Corresponding author.
E-mail addresses: fulianning@126.com (L. Fu), nanzou@sdu.edu.cn (N. Zou).
the beginning of the FLASHING DON’T WALK (FDW) signal phase.
Some studies have confirmed that pedestrian countdown signals
had generally positive effects on pedestrian behavior, because they
showed the time remaining which was helpful for crossing (Harre
and Wrapson, 2004; Eccles et al., 2004; Markowitz et al., 2006; Kim
et al., 2002; Xiong et al., 2014). It was precisely because of the
information showed by countdown signal, late starters increased
at the intersections with countdown signals (Hooper et al., 2007;
Wanty and Wilkie, 2010; Cleaver et al., 2011). Other researchers
found that the pedestrian countdown signals led to more pedestri-
ans crossing during the red light (York et al., 2011; Vujanić et al.,
2014). Meanwhile, other types of countdown signal timing were
also discussed, such as countdown running through the whole
walking interval including FDW interval or counted down FDW
interval only (Arhin and Noel, 2011), countdown terminating at
the beginning of yellow or beginning of red (Smith, 2014). But no
significant differences in pedestrian behavior and crash frequency
were found between different timings. Different from the types
mentioned above, some pedestrian countdown signals installed
in China counted down both during Red Man phase and Flashing

http://dx.doi.org/10.1016/j.aap.2016.05.017
0001-4575/© 2016 Elsevier Ltd. All rights reserved.
74 L. Fu, N. Zou / Accident Analysis and Prevention 94 (2016) 73–79
Fig. 1. Illustration of pedestrian signals at the two school intersections.
Green Man phase. This type of pedestrian countdown signals will
be discussed in the present study.
In addition to the pedestrians’ behavior, the pedestrian-vehicle
conflicts also need to be studied to evaluate pedestrian safety at the
intersections. Because conflicts were considered as good surrogates
for pedestrian collisions, and the conflict analysis provided more
information about the causes (Migletz et al., 1985; Garder, 1989).
Researchers found that the pedestrian-vehicle conflicts and crashes
decreased due to the installation of pedestrian countdown signals
(Eccles et al., 2004; Wanty and Wilkie, 2010; Huitema et al., 2014).
However, Richmond et al. (2014) found that pedestrian countdown
signals increased the pedestrian-motor vehicle collisions through
analyzing ten years collision data in Toronto. These indicated that
the pedestrian countdown signals might have different effects on
pedestrian-vehicle conflicts and crashes in different sites. Besides,
the characteristics of child pedestrian behavior were different from
adults. Therefore, it was necessary to carry out the conflicts study
to evaluate the effects of pedestrian countdown signals on children
crossing safety.
Despite some pedestrian countdown signal studies have men-
tioned children and youth (Supernak et al., 2013; Liao et al.,
2015), few studies have specifically targeting to the effects of
the countdown signals on children’s behavior and safety at the
school intersections. The main purpose of this study is to inves-
tigate the effects of pedestrian countdown signals on children’s
crossing behavior in school zones. Moreover, the conflicts between
child pedestrian and vehicles are also analyzed in this paper. And
then put forward some recommendations for improving children’s
crossing safety.
2. Method
The analysis in this study was based on the data obtained from
video recordings. Two primary schools were selected for inves-
tigation. The intersections in school zones had different types of
pedestrian signals, one was with traditional pedestrian signals and
the other one was installed with countdown signals. In China, the
pedestrian signals generally consisted of three symbols, namely the
Green Man (GM), Flashing Green Man (FGM) and Red Man (RM).
Both of the pedestrian signals in the two sites have these three
symbols, the difference lies in the countdown timer. Pedestrian
countdown signals can display the number of seconds remaining
until the end or beginning of the pedestrian green interval.
2.1. Sites description
The investigations were conducted in the city of Jinan, China.
The two intersections were near the schools, where most of the
children crossed the road. One school intersection was located at
Heping Road and Dianxindong Road (SI1), Heping Road was a bi-
directional and four lane road, expanding lanes each direction at
the intersection. The other school intersection was located at Wen-
huadong Road and Yangtouyudonggou Street (SI2). Wenhuadong
Road was also a two-way four-lane road, but without expanding at
the intersection. Dianxindong Road and Yangtouyudonggou Street
were both bi-directional road with two lanes. The vehicle flows at
these two school intersections were similar, 2064 vehicles per hour
at SI1 and 2112 vehicles per hour at SI2, respectively. Pedestrian
signal cycle length at SI1 was 180 s, the Green Man phase lasted
31 s, the Flashing Green Man phase was 10 s with counting down
from 9 to 0, the Red Man phase was 139 s. At the last 10 s of the Red
Man phase, the countdown timer displayed the remaining seconds
to the Green Man phase. The total pedestrian signal length at SI2
was 120 s, the Green Man phase was 22 s, the Flashing Green Man
phase was 5 s and the Red Man phase lasted 93 s. The illustration
of pedestrian signals at the two intersections were shown in Fig. 1.
2.2. Procedure
The observations were carried out in three weekdays under
good weather conditions. Child pedestrians were observed during
their school commute time. The three periods were as followed:
between 7:00 A.M. and 8:30 A.M. for the morning period, between
11:00 A.M. and 12:30 P.M. for the noon period and between
3:30 P.M. and 5:00 P.M. for the afternoon period. The cameras were
placed in the building near the intersection to make a better view of
child pedestrians and signal display, meanwhile child pedestrians
could not notice the camera.
Five types of information were extracted from the video record-
ings:
Type 1: Whether children starting to cross during the pedestrian
RM phase.
Type 2: Whether children starting to cross during the pedestrian
FGM phase.
Type 3: Whether children completing crossing during the GM
and FGM phases.
Type 4: Whether children running to cross.
Type 5: The conflicts between child pedestrians and motor vehi-
cles.
In this study, the conflicts were considered that when either
child pedestrians or motor vehicles took evasive action to prevent
collisions. Furthermore, the seconds of pedestrian started to cross
during the RM phase were calculated from the recordings. The gen-
der of children and other characteristics were also recorded, such
as whether accompanying with classmates or adults.
 L. Fu, N. Zou / Accident Analysis and Prevention 94 (2016) 73–79 75

Table 1
Descriptive statistics of the samples.

Description SI1 SI2 Responds

Gender of Children 53.6% (619) 52.4% (574) Male

46.4% (535) 47.6% (522) Female

Companions 31.3% (361) 37.0% (406) Adults

41.4% (478) 37.5% (411) Classmates
27.3% (315) 25.5% (279) Alone

Whether holding hands 60.4% (218) 72.2% (293) Yes

with adults 39.6% (143) 27.8% (113) No

Whether looking left and right 10.0% (115) 2.6% (28) Yes
before crossing 90.0% (1039) 97.4% (1068) No

Table 2
The impact of gender and companions on violation behavior.

Description Categories Cross during RM phase (%) Chi-square statistic Sig.

Gender Male 13.3 3.140 0.076

Female 10.9

Companions Adults 13.3 9.976 0.007

Classmates 9.6
Alone 14.6

2.3. Data analysis

The samples collecting from the two school intersections were

considered as two groups, and it assumed that they were inde-
pendent and random. The descriptive statistics were conducted
firstly, and then the two-tailed z-test at a 95% confidence level
was applied to examine whether there was difference of children’s
crossing behavior at intersections with countdown signals and
without countdown signals. The null hypothesis H0 and alternative
hypothesis H1 were as followed:

H0 : p1 = p2
(1)
H1 : p1 =
/ p2

where p1 was proportion of one type of crossing behavior at SI1 as

mentioned above, p2 was proportion of one type of crossing behav-
ior at SI2. Then the value of the test statistic was calculated by the
following formula: Fig. 2. Comparison of child pedestrian violation behavior.

p1 − p2
Z=  1  (2)
pc (1 − pc ) + 1
n1 n2
where n1 was the total number of pedestrians recorded at SI1, n2
was the total number of pedestrians recorded at SI2, pc was the
combined proportion of one type of crossing behavior.
p1 n1 + p2 n2
pc =
n1 + n2
If |z| > z˛/2 (1.96), the null hypothesis H0 would be rejected, it indi-
cated that there was no difference of child pedestrian crossing
behavior between the two school intersections.
3. Results
3.1. Descriptive statistics
The sample size was 1154 at SI1 and 1096 at SI2 respectively. The
results of descriptive statistical analysis were shown in Table 1. It
showed that the percentage of boys was higher than girls both at SI1
and SI2. Child pedestrians were divided into three groups accord-
ing to the companions: be accompanied by adults, be accompanied
with classmates, alone. The results indicated that more than 40%
of children were accompanied with classmates at SI1, and this pro-
portion at SI2 was 37.5%. The proportion of children crossing alone
was lower comparing to the other two groups. Among the adults,
the percentages of mother and grandmother were relatively higher.
In the group of children accompanied by adults, almost 40% chil-
dren did not hold hands with adult at SI1, whereas the proportion
at SI2 was lower (27.8%). It was important to look left and right
before crossing, but few children were observed to have this safety
(3)
behavior at both sites.
3.2. Children start to cross during the RM phase
The number of child pedestrian who started crossing during the
RM phase were counted. To investigate the effects of countdown
signal timers on child pedestrian violation behavior, the propor-
tion of child pedestrian violations were computed and z-test was
used to compare between the two school intersections (see Fig. 2).
At SI1, 20.9% of child pedestrians crossed against the red light, it
was obviously higher than the proportion at SI2. The z-test result
showed significant difference of child pedestrian starting to cross
during the RM phase between the two intersections (Z = 12.958,
p < 0.05).
Further analyses were conducted to understand whether gen-
der and companions influenced child pedestrian violation crossing
76 L. Fu, N. Zou / Accident Analysis and Prevention 94 (2016) 73–79

Fig. 4. Comparison of completing crossing behavior during GM and FGM phases.

Fig. 3. Comparison of crossing behavior during FGM phase.

3.5. Children run to cross

behavior, the results were shown in Table 2.The violation propor-
tion of boys was higher than girls, but the result was not significant
at the 0.05 level. Children who crossed with classmates together
had lowest proportion of violation. By contrast, children who were
accompanied by adults or alone to cross had higher proportion of
violation. The differences between different companion groups had
a 0.007 level of significance.
3.3. Children start to cross during the FGM phase
Children who started to cross when the pedestrian green man
had been flashing were recorded. The two school intersections
in this study had different FGM phase. Countdown timers were
installed at SI1 to display the remaining time to end the green light,
but no countdown timers at SI2. The results showed that the pro-
portions of child pedestrian crossing during FGM phase at SI1(6.9%)
and SI2(6.2%) were similar (see Fig. 3), z-test results also showed
that there was no significant difference in children starting to cross
during the FGM phase between SI1 and SI2 (Z = 0.696, p > 0.05).
The Chi-square test results showed that there was also no sig-
nificant difference between boys and girls in “adventurer”, those
who starting to cross during the pedestrian FGM phase (Table 3).
But significant differences were found in the case of adventure
crossing behavior between different companion groups. Children
who crossed with classmates had significant lower proportion of
adventure behavior than the other groups.
Many children had running behavior during crossing the road.
Apart from children were physically active who might be more
likely to running, countdown signal timers might also have effects
on their behavior. Children who ran to cross were recorded during
data collections. The analysis results showed that a larger propor-
tion of children ran to cross at SI1 with countdown timers than SI2
without countdown. The z-test results indicated that the difference
was significant at the 0.05 level (Z = 4.009, p < 0.05).
More boys were observed running to cross than girls in this
study. The Chi-square test results showed that the difference was
significant (see Table 5). Furthermore, children who crossed alone
had a significant higher proportion of running behavior than chil-
dren were accompanied by adults or with classmates (Fig. 5).
3.6. Time interval distribution of children violations
The Red Man phase was divided into different intervals of 5 s
with the purpose of studying the distribution of child pedestrian
violations. The time distribution analysis results of children viola-
tors showed that there was a small number peak of child pedestrian
starting to cross at the first 5 s of the RM phase, both at SI1 and SI2.
The number of children violations during the initial RM interval at
SI1 was larger than that at SI2. During the intermediate of the RM
phase, the number of violations at SI1 was also larger than SI2. Fur-
thermore, the major violations were found in the last 10 s of RM
phase at SI1, especially in the last 5 s, the violation number comes

3.4. Children complete crossing during GM and FGM phases

The safety of crossing would reduce if child pedestrians did not

complete crossing before red light onset. Fig. 4 showed that the
proportion of children not completing crossing during GM and FGM
phases at SI2 (20.3%) was higher than it was at SI1 (14.0%). The z-
test results indicated that there was significant difference of not
completing crossing behavior during GM and FGM phases between
SI1 and SI2 (Z = −3.977, p < 0.05).
The analysis results of different companion categories in Table 4
indicated that there were significant differences in the proportion
of not completing crossing behavior before red light. More chil-
dren crossed with classmates finished their crossings during GM
and FGM phases than children who were accompanied by adults or
alone. However, no significant difference was found in this type of
behavior between boys and girls. Fig. 5. Comparison of running to crossing behavior.
 L. Fu, N. Zou / Accident Analysis and Prevention 94 (2016) 73–79 77

Table 3
The impact of gender and companions on starting to cross behavior during FGM phase.

Description Categories Cross during FGM phase (%) Chi-square statistic Sig.

Gender Male 6.0 1.621 0.203

Female 7.3

Companions Adults 8.1 29.647 0.000

Classmates 3.1
Alone 9.8

Table 4
The impact of gender and companions on not completing crossing behavior during GM and FGM phases.

Description Categories Not completing crossing during GM and FGM phases (%) Chi-square statistic Sig.

Gender Male 18.1 2.110 0.146

Female 15.8

Companions Adults 17.2 6.808 0.033

Classmates 14.8
Alone 20.0

Table 5
The impact of gender and companions on running to cross behavior.

Description Categories Cross during RM phase (%) Chi-square statistic Sig.

Gender Male 23.6 5.415 0.020

Female 19.6

Companions Adults 12.6 167.469 0.000

Classmates 17.2
Alone 40.2

Table 6
Conflicts between child pedestrian and straight through vehicles.

The reasons of child pedestrian-vehicle conflicts Straight through vehicles

SI1 SI2

Child pedestrian crossing against the red light 18 4

Child pedestrian walking out of the crosswalk 9 1
Child pedestrian did not complete crossing before the red light onset 0 3
The vehicles did not leave intersection before the red light onset 0 7

Total 27 15

to the highest peak. By contrast, there is almost no violation during sive left or right turning phases for vehicles was the main reason
the intermediate and final of the RM phase at SI2 (Fig. 6). causing the conflicts between child pedestrian and turning vehicles.

3.7. Child pedestrian-vehicle conflicts
The child pedestrian-vehicle conflicts were observed and
counted, the conflicts between different vehicles and the conflicts
caused by red-driving vehicles were not included in this study.
Child pedestrian conflicts between straight through vehicles and
turning vehicles were recorded separately. The results in Table 6
showed that there were more child pedestrian-straight through
vehicle conflicts at SI1 than SI2 overall. At SI1, conflicts between
child pedestrian and straight through vehicles were caused by two
reasons: one was that children crossed against the red light and
the other one was that children did not walk along the crosswalk.
Two thirds of the pedestrian-straight through vehicle conflicts
occurred due to children did not comply with signal, crossing dur-
ing the Red Man phase. By contrast, less children crossing against
the red light leading to conflicts with straight through vehicles at
SI2. However, more conflicts occurred at SI2 because vehicles did
not leave intersection before red light onset and children did not
complete crossing before red light onset. Child pedestrian-turning
vehicle conflicts analysis results indicated that this type of conflicts
occurred frequently too. The number of child pedestrian-turning
vehicles conflicts were 21 at SI1 and 17 at SI2 respectively. No exclu-
4. Discussion
4.1. Key findings in this study
The results of the present analysis have shown that children’s
crossing behavior are affected by the countdown signals. One of
the main outcomes is that countdown signals make higher propor-
tion of children violators during RM phase, compared to without
countdown signals. Further analysis of time distribution of child
pedestrian violations indicates that the violations at SI1 were
occurred mainly in the last 5 s of the RM phase. One possible rea-
son was that there were pedestrian countdown timers at SI1, which
could display the number of seconds remaining to the GM phase.
Children are more impulsive and impatient, therefore, many of chil-
dren cannot waiting for the pedestrian green light to start crossing.
In addition, many running children were found during the last
seconds of RM phase. On the one hand, children know the pedes-
trian green light was coming through reading the countdown timer.
On the other hand, quickly crossing to avoid vehicles still on the
road was needed once they decided to start crossing during the
RM phase. Consequently, the countdown starting at the final of the
RM phase have negative effects on children’s crossing behavior. It
may cause children’ risky crossing behavior, thereby reducing the
78 L. Fu, N. Zou / Accident Analysis and Prevention 94 (2016) 73–79
Fig. 6. Distribution of child pedestrian violations during the RM phase.
safety of crossing. Previous study has shown that the probability
of pedestrian accidents is much higher when crossing at the final
moments before the pedestrian green light is on (Lipovac et al.,
2013). So, countdown display during the RM phase is not suggested
to be applied at school intersection. Furthermore, the violation pro-
portion of children crossing with classmates is relatively lower than
children crossing accompanying by adults or alone. Other studies
also proved that pedestrians tended to start crossing during the red
light when being alone (Keegan and O’Mahony, 2003; Rosenbloom,
2009; Vujanić et al., 2014).
Another important finding is that comparing to the school
intersection without countdown signals (SI2), more child pedes-
trians can complete crossing during the green light phase at school
intersection with countdown signals (SI1). There are two possible
explanations: one is due to SI1 has countdown timer to display the
remaining seconds till next red light onset, it’s helpful for children
to adjust their crossing speeds, avoid getting caught in the middle
of crosswalk; on the other hand, the green man flashing time at
SI1 is longer than SI2, five seconds is too short as reminder time
for child pedestrian to complete crossings. Thus, how to design and
optimize the length of time to countdown during the FGM phase
require further study in future. However, the results suggest that
countdown signals have no significant effect on children starting
to cross during the FGM phase.
The conflicts analysis results showed that the lack of compliance
led to more child pedestrians exposed to straight though vehicles
and increased the risk at school intersection. More children were
found to start crossing during the last 5 s of the RM phase at SI1
due to the countdown started at the end of RM phase, these chil-
dren might create more conflicts with vehicles. Because there were
still vehicles passing through the intersection in this period and
even with higher speed. It was proved that more pedestrian-vehicle
conflicts were caused by these children violators at SI1. Then the
safety of crossing was reduced as the increase of conflicts. So, it
was not a good choice to set the countdown at the end of RM phase.
There were also many conflicts between children and turning vehi-
cles because both SI1 and SI2 had no exclusive left or right turning
phases. The results above had shown that more children could not
finish crossing before the red light displayed at SI2, some conflicts
were observed between these children and vehicles at this inter-
section. But no pedestrian-vehicle conflicts occurred on account of
children not finishing crossing at SI1. In this regard, the countdown
signal which displayed during FGM phase had positive effect on the
improvement of children’s crossing safety.
Besides, gender and companion factors were also examined in
this study. The influences of companions were significant on dif-
ferent children crossing behavior. In summary, children who were
alone crossing the road had more risk-taking crossing behavior
than children who had companions. Especially children who were
accompanied with classmates had less risk-taking behavior. This is
probably because they talk to each other when crossing in company
with classmates, they are less likely to break talking to take risky
behavior such as running. Besides, children who has higher risk per-
ception may influence the companions. But this does not mean that
children crossing the road with classmates is safer. Because they
will get distracted if they talk to each other when crossing the road,
causing no observation of surrounding traffic conditions. So chil-
dren’s traffic safety consciousness still need to be improved through
different intervention measures. The gender difference was not sig-
nificant in this study, only found that boys were more likely to run
to cross the road.
4.2. Recommendations
Based on the results in this study, different countermeasures
are proposed to improve child pedestrian safety at school inter-
sections. On the one hand is the improvement of facilities, such as
the implementation of pedestrian countdown signals. The pedes-
trian countdown signals are only suggested to display during the
FGM phase, while not during the RM phase. In addition, it is also
required to further study on the length of countdown timing dur-
ing FGM phase. In the aspect of signal timing, a separate pedestrian
green phase is recommended to set at school intersections, such as
scramble period. During the scramble period pedestrians in all the
crosswalks can cross and also diagonally across are allowed. All the
vehicles are stopped by the red signals during this period. It can
help to reduce the conflicts between child pedestrians and vehi-
cles. But setting scramble period may increase the delay of vehicles
at the intersection. So, different signal timing plans are suggested
depending on the school commute time in a day. Furthermore, the
length of pedestrian green phase should be longer during school
commute time, considering that there are many older pedestrians
who pick up children crossing at school intersections. Besides the
optimization of signalization, better geometry and facility designs
should be practiced at school intersections. For example, diago-
 L. Fu, N. Zou / Accident Analysis and Prevention 94 (2016) 73–79
nal crosswalks are recommended to design at school intersections.
Children and parents were observed to cross diagonally for conve-
nience in this study. Therefore, diagonal crosswalks are also needed
corresponding the scramble period. In addition, other reasonable
design should also be given attention to improve child pedestrian
safety, such as crosswalk width, refuge island, application of intel-
ligent facilities etc. Finally, awareness and education are critical to
changing children’s risky behavior and preventing the accidents.
School and other safety relevant departments should co-operate to
carry out more programs or classes to hence their risk awareness.
4.3. Limitations and further research
Overall, this study has investigated the effects of countdown
signals on child pedestrians and obtained some meaningful results.
However, this study has some limitations in the design and method-
ology. The observations were conducted at two school intersections
in a city, and carried out in three days, which could not represent all
the situations and the changes over time. In addition, some other
factors that may influence children’s behavior and child pedestrian-
vehicles conflicts were overlooked in this research, such as vehicle
speed, length and width of crosswalk and other geometrical designs
of the intersection. Although two similar school intersections were
selected to minimize the influences of these factors. So, long-term
and larger scale observations studies are required to conduct to
validate the findings in this study. And a before-after study on this
issue is desired to be carried out cooperating with related depart-
ments in the next research. Finally, the effects of different types of
pedestrian countdown signals on children’s crossing behavior and
safety should be explored in the future.
5. Conclusion
To sum up, the results in this study show different effects of
countdown signals on children’s crossing behavior. Setting count-
down at the last moment of the Red Man phase lead to the increase
of children violation behavior. And these violators create more
conflicts with vehicles. Consequently, removing countdown at the
end of the Red Man phase may improve child pedestrian’s cross-
ing behavior and reduce the conflicts with vehicles. Nevertheless,
countdown during Flashing Green Man are effective at helping chil-
dren to completing crossing, avoid getting caught in the crosswalk
before the Red Man onset. Moreover, children who cross the road
alone have more risk-taking behavior than children have com-
panions. Finally, the findings in this study may provide helps for
designers and policy makers how to design the pedestrian count-
down signals and improve children’s crossing safety in school
zones.
References
Arhin, S.A., Noel, E.C., 2011. Evaluation of the impact of two countdown pedestrian
signal displays on pedestrian behavior in an urban area. Int. Conf. Saf. Secur.
Eng. 4th, http://dx.doi.org/10.2495/safe110311.
79
Cleaver, M.A., Hislop, J., de Roos, M.P., Fernades, R., Prendergast, M., Brisbane, G.,
Levasseur, M., McTiernan, D., 2011. An evaluation of pedestrian countdown
timers in the Sydney CBD. Australasian Road Safety Research Policing and
Education Conference, 1–9.
Eccles, K., Tao, R., Mangum, B., 2004. Evaluation of pedestrian countdown signals in
montgomery county, Maryland. Transp. Res. Rec. 1878, 36–41, http://dx.doi.
org/10.3141/1878-05.
Garder, P.E.R., 1989. Pedestrian safety At traffic signals: a study carried out with
the help of a traffic conflicts technique. Accid. Anal. Prev. 21, 435–444.
Harre, N., Wrapson, W., 2004. The evaluation of a central-city pedestrian safety
campaign. Transp. Res. Part F Traffic Psychol. Behav. 7, 167–179, http://dx.doi.
org/10.1016/j.trf.2004.07.002.
Hooper, M., Vencatachellum, V., Tse, M., 2007. Trial of pedestrian signals
incorporating a numerical countdown display in Auckland CBD. In: IPENZ
Transportation Group Conference, Tauranga.
Huitema, B.E., Van Houten, R., Manal, H., 2014. Time-series intervention analysis of
pedestrian countdown timer effects. Accid. Anal. Prev. 72, 23–31, http://dx.doi.
org/10.1016/j.aap.2014.05.025.
Keegan, O., O’Mahony, M., 2003. Modifying pedestrian behaviour. Transp. Res. Part
A Policy Pract. 37, 889–901, http://dx.doi.org/10.1016/S0965-8564(03)00061-
2.
Kim, K.W., Kim, Y., Seo, H.Y., 2002. An evaluation of pedestrian countdown signals.
KSCE J. Civ. Eng. 6, 533–537.
Lambrianidou, P., Basbas, S., Politis, I., 2013. Can pedestrians’ crossing countdown
signal timers promote green and safe mobility? Sustain. Cities Soc. 6, 33–39,
http://dx.doi.org/10.1016/j.scs.2012.07.005.
Liao, D., Yu, B., Ma, W., 2015. Empirical analysis of countdown signals on pedestrian
behaviour. Transport 168, 15–22, http://dx.doi.org/10.1680/tran.11.00078.
Lipovac, K., Vujanic, M., Maric, B., Nesic, M., 2013. The influence of a pedestrian
countdown display on pedestrian behavior at signalized pedestrian crossings.
Transp. Res. Part F Traffic Psychol. Behav. 20, 121–134, http://dx.doi.org/10.
1016/j.trf.2013.07.002.
Markowitz, F., Sciortino, S., Fleck, J.L., Yee, B.M., 2006. Pedestrian countdown
signals: experience with an extensive pilot installation. ITE J. Inst. Transp. Eng.
76, 43–48.
Migletz, D.J., Glauz, W.D., Bauer, K.M., 1985. Relationships Between Traffic
Conflicts and Accidents. Federal Highway Administration, Washington, DC.
Richmond, S.A., Willan, A.R., Rothman, L., Camden, A., Buliung, R., Macarthur, C.,
Howard, A., 2014. The impact of pedestrian countdown signals on
pedestrian-motor vehicle collisions: a reanalysis of data from a
quasi-experimental study. Inj. Prev. 20, 155–158, http://dx.doi.org/10.1136/
injuryprev-2012-040717.
Rosenbloom, T., 2009. Crossing at a red light: behaviour of individuals and groups.
Transp. Res. Part F Traffic Psychol. Behav. 12, 389–394, http://dx.doi.org/10.
1016/j.trf.2009.05.002.
Smith, T.B., 2014. Experimentation With Countdown Pedestrian Signal Timing and
The Effect on Intersection Safety. University of Delaware in Partial.
Supernak, J., Verma, V., Supernak, I., 2013. Pedestrian countdown signals: what
impact on safe crossing? Open J. Civ. Eng. 3, 39–45.
Vujanić, M., Pešić, D., Antić, B., Smailović, E., 2014. Pedestrian risk At the signalized
pedestrian crossing equipped with countdown display. Int. J. Traffic Transp.
Eng. 4, 52–61, http://dx.doi.org/10.7708/ijtte.2014.4(1).04.
Wanty, D.K., Wilkie, S.M., 2010. Trialling Pedestrian Countdown Timers at Traffic
Signals NZ. Transport Agency, Wellington.
Xiong, H., Xiong, L., Deng, X., Wang, W., 2014. Evaluation of the impact of
pedestrian countdown signals on crossing behavior. Adv. Mech. Eng. 2014,
1–7, http://dx.doi.org/10.1155/2014/518295.
York, I., Ball, S., Beesley, R., Webster, D., Knight, P., Hopkin, J., 2011. Pedestrian
Countdown at Traffic Signal Junctions (PCaTS)—Road Trial. Transport for
London.