Content uploaded by Mark Johann King
Author content
All content in this area was uploaded by Mark Johann King
Content may be subject to copyright.
Journal of the Australasian College of Road Safety – August 2010
56
Abstract
Visibility limitations make cycling at night particularly
dangerous. We previously reported cyclists’ perceptions of their
own visibility at night and identified clothing configurations
that made them feel visible. In this study we sought to
determine whether these self-perceptions reflect actual visibility
when wearing these clothing configurations. In a closed-road
driving environment, cyclists wore black clothing, a fluorescent
vest, a reflective vest, or a reflective vest plus ankle and knee
reflectors. Drivers recognised more cyclists wearing the
reflective vest plus reflectors (90%) than the reflective vest alone
(50%), fluorescent vest (15%) or black clothing (2%). Older
drivers recognised the cyclists less often than younger drivers
(51% vs 27%). The findings suggest that reflective ankle and
knee markings are particularly valuable at night, while
fluorescent clothing is not. Cyclists wearing fluorescent clothing
may be at particular risk if they incorrectly believe themselves to
be conspicuous to drivers at night.
Keywords
Night visibility, Cyclists, Reflective clothing, Age
Introduction
Cyclists are considered to be among the most vulnerable of all
road users. They have among the largest proportion of self-
reported near-miss crashes, significantly higher than that of
motorists and comparable to that of pedestrians [1]. The injury
consequences of a crash are also more severe for cyclists, where
the probability of a cyclist being seriously injured following
involvement in a crash was found to be almost 27% in
Australian data collected over a four-year period [2]. The
vulnerability of cyclists was further highlighted by Sonkin et al.
[3], who reported that while child pedestrian fatality rates per
10 million miles fell from 1.08 to 0.27 (75%) during the
period 1985 to 2003, child cyclist fatality rates only decreased
from 0.84 to 0.55 (35%) per 10 million miles travelled.
Night-time cycling has been shown to be more dangerous than
cycling in daylight, with 40% of cyclist fatalities occurring at
night despite much lower exposure rates than in the daytime
[4]. Rodgers [5] notes that while only 12% of cyclists reported
that they rode after dark, 35% of cyclist deaths occur outside of
daylight hours. The role of visibility in contributing to fatal
accidents was examined further by Owens and Sivak [6], who
found that 78.8% of all fatal collisions involving vulnerable
road users (cyclists or pedestrians) occurred during low-light
conditions. When visibility was degraded further by poor
atmospheric conditions, such as rain or fog, 92.3% of all fatal
accidents involving a vulnerable road user occurred in low-light
conditions [6]. A high proportion of cyclist fatalities have been
reported to be related to problems with frontal rather than rear
conspicuity [7], and motorists involved in night-time collisions
with cyclists commonly report that they did not see the cyclist
until it was too late to avoid a collision [8, 9].
The use of static or flashing front and rear bicycle lights is one
widely adopted approach for improving cyclist visibility and is
now a legal requirement when cycling on roads at night in
many countries, including Australia. Another relatively
inexpensive and practical approach to improving the conspicuity
of cyclists is the use of high-visibility clothing.
In a previously published survey of 1460 participants (622
drivers and 838 cyclists), Wood et al. [10] explored the beliefs
and attitudes of cyclists and drivers regarding cyclist visibility
and safety, and cyclists’ use of different clothing configurations,
with a particular focus on improving visibility under reduced
illumination conditions, including dawn, dusk and night-time.
In that study we found that cyclists believe they are more visible
(and that they are visible at longer distances) than did drivers
under the same circumstances.
This was early evidence that, like pedestrians [11], cyclists may
overestimate their own visibility in low light conditions. The
survey also revealed that although cyclists endorsed the use of
high-visibility clothing and aids, particularly in low-light
conditions, relatively few cyclists reported wearing selected
high-visibility clothing on a regular basis. Cyclists as a group
may thus underestimate the importance of attracting other road
users’ attention when visibility is limited, such as under night-
time conditions.
In the study described above [10], cyclists also rated wearing a
reflective vest as being the most effective means of improving
visibility, over and above the use of reflective strips worn on the
moveable joints. This is relevant because empirical research on
the night-time conspicuity of pedestrians has repeatedly
revealed the opposite: that reflective strips on the major
moveable joints are highly effective in improving pedestrian
conspicuity, presumably due to humans’ high perceptual
sensitivity to distinctively human patterns of joint movement
(‘biological motion’ or ‘biomotion’) [12].
Cyclist visibility at night: Perceptions of visibility do
not necessarily match reality
by JM Wood*, RA Tyrrell**, R Marszalek*, P Lacherez*, T Carberry*, BS Chu*, and MJ King***
*School of Optometry and Institute of Health and Biomedical Innovation, Queensland University of Technology,
Brisbane, Australia
**Department of Psychology, Clemson University, South Carolina, USA
***Centre for Accident Research and Road Safety - Queensland (CARRS-Q), Queensland University of Technology,
Brisbane, Australia
ACRS Journal 21_No3:ACRS Journal Vol 17 No 2 10/8/10 2:06 PM Page 56
Journal of the Australasian College of Road Safety – August 2010
57
Numerous studies have demonstrated that drivers are able to
recognise the presence of pedestrians more often and at much
longer distances when they are wearing reflective strips in a
biomotion configuration, than when they wear a reflective vest
[13-15]. It is thought that reflective vests are less useful because
they limit the placement of the reflective material to the torso,
which presents much less motion information to approaching
drivers. Although the patterns of movement involved in cycling are
inherently different from those associated with being a pedestrian,
highlighting a cyclist’s movements (by placing reflective markings
on the cyclist’s ankles and knees) might be an effective low-cost
approach to enhancing cyclist conspicuity.
In our survey [10] we also found that cyclists may overestimate the
usefulness of some visibility aids – for example, fluorescent clothing
– at night. Fluorescent clothing acts by converting the wavelength
of ultra-violet (UV) light (present in sunlight) to longer visible
wavelengths, which leads to an overall increase in reflected visible
light under daytime conditions. However, streetlights and vehicle
headlights do not provide substantial amounts of UV; thus,
fluorescent materials are not a particularly valuable conspicuity aid
during typical night-time conditions. Interestingly, the majority of
the cyclists and drivers in our survey considered fluorescent bicycle
clothing to be more visible at night than white clothing. Therefore,
road users may also be inadequately informed regarding the
limitations of certain visibility aids. The failure of road users to
understand such issues could be critical.
In the current study we evaluated the benefits of a range of
visibility aids for cyclists under real world night driving conditions.
These data are important, as without objective evidence
demonstrating the effect of improving visibility on drivers’
perceptions and reactions to cyclists on the road, it is not possible
to inform cyclists or other road users with regard to their benefits,
or indeed possible limitations.
We included both young and older drivers in this study in order to
explore the extent to which driver age impacts on night-time cyclist
visibility, given that previous studies have shown that pedestrian
visibility at night is significantly impaired with increasing age [13,
14]. We compared the on-road data collected here with the
perceptions of cyclists’ own visibility that we had gathered in our
previous survey-based study, which determined how well cyclists’
perceptions of visibility aids aligns with the actual benefits of
visibility aids [10].
Methods
In this study volunteer participants drove around a closed road
driving circuit at night and indicated when they recognised the
presence of a cyclist wearing a range of different clothing
configurations.
Participants
Participants included 12 young (M = 25.3 years, range 18-35)
and 12 older (M = 72.5 years, range 66-80) visually normal
individuals who had a current driver’s licence and were regular
drivers, with a visual acuity of 6/9 (20/30) or better. The study
was conducted in accordance with the requirements of the
Queensland University of Technology Human Research Ethics
Committee.
Closed-road test circuit and experimental vehicle
All driving was conducted under night-time conditions and was
assessed on a 1.8km closed-road circuit [13, 14]. The circuit,
which is representative of a rural road, consisted of a two- to
three-lane bitumen road and included hills, bends, curves, lengthy
straight sections, and standard road signs and lane markings.
There was no additional ambient lighting on the circuit, and
experimental sessions were only conducted on nights when there
was no rain and the road surface was dry.
Two cyclists were positioned at different locations around the
circuit, and pedalled in place on a resistance trainer so as to ensure
naturalistic cyclist motion, while maintaining a consistent location
that is critical for purposes of experimental control (Figure 1).
Each cyclist was equipped with a two-way radio, as was the
experimenter in the test vehicle. This allowed all communications
regarding participant clothing to be conducted between laps and
outside of the vehicle, so that the participants could not hear the
conversations.
The data presented in this paper relate to the test cyclist positioned
at point ‘A’ at the top left of Figure 1. In order to isolate the effects
of clothing on cyclist visibility, the bicycle did not have front or
rear lights. We have previously observed that a significant
proportion of cyclists do not always use their lights under low-
light conditions, and so this reflects a reality of night driving [10].
Figure 1. Schematic map of the closed-road circuit showing the location of the two cyclists and the three clutter zones.
The test vehicle’s direction of travel is indicated by arrows. The position of the glare lights are indicated by stars.
ACRS Journal 21_No3:ACRS Journal Vol 17 No 2 10/8/10 2:06 PM Page 57
Journal of the Australasian College of Road Safety – August 2010
58
To simulate the effects of other vehicles being present, two pairs
of battery-powered headlights were placed along the circuit, in
close proximity to the test cyclist. The glare lights were
positioned 5.4m in front of and 5.1m to the left of the test
cyclist (when viewed from the perspective of an approaching
driver). To the driver, these headlights approximated an
oncoming vehicle’s headlights and the glare from them added a
degree of visual challenge. The test cyclist was positioned in the
outermost oncoming lane (as seen from the participant’s point
of view) and was a minimum of two lanes removed from the
test vehicle. The test cyclist was also surrounded by clutter
provided by an array of reflective cones and posts.
To provide an additional degree of visual complexity and also to
act as distracters, three additional clutter zones were set up
along the circuit. Two of the clutter zones consisted of small-
and medium-sized retro-reflective traffic cones, large retro-
reflective posts and flashing amber lights. The third zone
consisted solely of three pairs of large retro-reflective traffic
cones and was used as a ‘navigation zone’ – where the driver
was required to guide the test vehicle through the zone without
hitting any cones. This was done to increase driver workload.
White flashing LEDs were positioned at three locations around
the circuit, which served to simulate an oncoming bicycle and
also to reduce the expectancy of the drivers. The LEDs were
positioned on the right-hand shoulder of the road on black
posts at a height approximating the front light of a bicycle.
The direction that the test cyclist faced was also varied between
laps to simulate the two most common crash configurations
reported by cyclists in our previous paper [10], where 38%
reported a crash in which the motorist collided with a cyclist
turning across their path when they were both heading in the
same direction, and 19% reported being sideswiped. For half of
the laps the cyclist faced in the same direction as the driver, and
for half the laps the cyclist was positioned side-on to the driver
(see Figure 2), as if they were about to enter the traffic from a
street at 90 degrees to the driver’s direction of travel.
Figure 2. Photograph of the section of the road circuit
where Cyclist A was positioned
Clothing conditions
For each lap, the test cyclist wore one of four clothing outfits:
(1) a black tracksuit, (2) a black tracksuit with a fluorescent
yellow cycling vest with no retro-reflective materials present, (3)
a black tracksuit plus a fluorescent yellow cycling vest that
included silver retro-reflective markings (Netti Litehook®) on
the shoulders, front and rear of the torso, or (4) the same black
tracksuit and retro-reflective vest with the addition of 50mm-
wide silver retro-reflective strips (3M Scotchlite® 8910 silver
fabric) positioned on the cyclist’s ankles and knees.
Asecond cyclist was present on all laps at location B (see Figure
1). This reduced the participants’ ability to associate a cyclist
with a particular location on the circuit. The second cyclist
wore the same range of clothing configurations as the test
cyclist in an independently determined random order (minus
the fluorescent vest).
Procedures
Participants drove around the circuit in a right-hand drive sedan
fitted with two digital video cameras mounted on the roof of
the vehicle [16]. The system recorded two overlapping images
of the road scene and was linked to a LED marking system,
which recorded the moment the participant pressed a large
luminous dash-mounted touch pad to indicate recognition.
Participants were given a practice lap in order to familiarise
themselves with the car, the road circuit and the tasks required
of them. The practice lap was followed by 10 data collection
laps. These comprised the eight combinations of cyclist clothing
and bicycle direction presented in a random order for each
participant, plus two laps where the test cyclist was absent,
which was held constant between participants.
Participants were instructed to follow the specified route, to
drive at a comfortable speed and to press the touchpad
whenever they recognised that a cyclist was present in the road
scene ahead. Participants were instructed to read aloud all road
signs encountered so as to increase driver workload (these data
were not recorded). To quantify the participants’ responses to
the cyclist we recorded whether the participant pressed the
response button at any point along their approach to the cyclist.
Thus, we could track the percentage of trials in which the
participants correctly identified the presence of the test cyclist.
Results
An independent samples t-test was conducted on the
proportion of cyclists recognised by each participant, according
to the age group of the participants. There were four false
sightings of pedestrians over the total of 240 laps, but this
number was too small to be usable in the analysis.
Overall, younger drivers identified nearly twice the number of
cyclists as did older drivers; on average, younger drivers
identified just over half of the cyclists (51%), whereas older
drivers identified just over a quarter (27%). This age effect was
ACRS Journal 21_No3:ACRS Journal Vol 17 No 2 10/8/10 2:06 PM Page 58
Journal of the Australasian College of Road Safety – August 2010
59
significant, t(22) = 4.12, p < .001. As can be seen in Figure 3,
older drivers did not detect any of the cyclists wearing black or
fluorescent clothing, and less than half of the cyclists wearing
reflective vests. Younger drivers performed much better;
however, they detected less than half of the cyclists wearing
black or fluorescent clothing.
A
B
Figure 3. Percentage of cyclists recognised as a function of
clothing and cyclist direction for the young (A) and older
(B) drivers
A repeated measures t-test was conducted on the proportion of
cyclists recognised by each participant according to whether the
cyclists were entering the roadway as if from a side road or
were pedalling in the same direction as the driver, and found no
significant differences t(22) = .81, p = .427. Data were thus
combined across the direction conditions, to enable a two-way
analysis of variance between clothing and age in terms of the
number of cyclists correctly identified.
The analysis revealed a large overall effect of clothing, F(3,66)
= 45.7, p < .001. Overall, drivers identified the largest number
of cyclists wearing the vest plus the ankle and knee reflectors
(90% correctly recognised), followed by the reflective vest alone
(50%), the fluorescent clothing (15%), and lastly black clothing
(2%). All pair-wise differences were significant, with the
exception of black and fluorescent clothing, which were not
significantly different from one another.
There was no significant interaction between age and clothing,
indicating that the effects of clothing were similar for the two
age groups, F(3,66) = 1.83, p = .151. While older drivers
were less likely than young drivers to identify the cyclists, the
degree to which they were less successful than young drivers did
not vary according to clothing configuration.
Discussion
In this field study we sought to determine how the ability of
drivers of different ages to recognise the presence of cyclists at
night-time is influenced by the cyclist’s clothing. The data
demonstrate that cyclist clothing and driver age both
significantly affect the ability of drivers to recognise cyclists
under real world night-time driving conditions. Collectively
these results are important, particularly when considered in the
context of our previously collected data regarding cyclists’
perceptions of their own visibility and how often they wear
such visibility aids.
There was a strong effect of clothing on the percentage of
cyclists who were recognised by drivers. Adding ankle and knee
markings to a typical reflective cycling vest provides a powerful
enhancement of the cyclist’s conspicuity. This manipulation
increased the percentage of drivers who recognised that a cyclist
was present from 50% to 90% overall, with 100% of cyclists
being recognised by the younger cohort of drivers.
Even though this configuration did not use a ‘full’ biological
motion configuration, the effect was just as robust as those
demonstrated in prior studies for pedestrian visibility [13-15].
That the cyclist only wore the reflectors on the ankles and knees
and yet was still easily recognised suggests that ‘full’ biological
motion (i.e., placing reflectors on all major moveable joints)
may not be necessary for the successful recognition of cyclists,
and that a convenient subset – marking just the ankles and
knees – may be sufficient. This hypothesis will be further
explored in our future studies.
Recognition of the cyclist wearing the reflective vest without
the ankle and knee markings (50%) was better than for the
cyclist wearing either the fluorescent (15%) or black clothing
(2%), but for older drivers, recognition levels for the vest were
as low as 30%. This may be a surprise to the many cyclists who
rely on reflective vests as a visibility aid at night. It may not,
however, be unexpected to researchers, who have previously
demonstrated that pedestrians also have a strong tendency to
overestimate their own visibility at night and to underestimate
the conspicuity benefits provided by biological motion [11].
The relatively low conspicuity levels of the cyclists when
wearing the reflective vest alone is likely to be attributable to
the lack of perceptible torso motion signifying the presence of a
cyclist. Importantly, in our survey [10], cyclists ranked reflective
vests as being most visible under reduced illumination
ACRS Journal 21_No3:ACRS Journal Vol 17 No 2 10/8/10 2:06 PM Page 59
Journal of the Australasian College of Road Safety – August 2010
60
conditions. Thus, typical cyclists do not seem to appreciate that
reflective vests may not maximise their conspicuity and that
biological motion markers can increase their conspicuity. Added
to this is the low level of use of visibility aids in general by
cyclists at night, with only 35% of cyclists reporting that they
wear reflective clothing either ‘often’ or ‘always’.
Our findings that the visibility benefits of fluorescent vests are
small, that they do not offer a significant improvement on black
clothing, and that older drivers fail to recognise cyclists wearing
fluorescent clothing on any trials are important when
considered in light of the results of our earlier survey [10]. In
that study both cyclists and drivers rated the visibility benefits
of fluorescent vests to be high even under night-time
conditions; indeed, there was little difference in their ranking of
the visibility benefits of the fluorescent clothing for either day
or night-time conditions.
However, fluorescent materials have little visibility benefit at
night, as they are activated only by UV radiation, which is
lacking in headlights and streetlights. Cyclists appear to assume
incorrectly that the visibility advantage of fluorescent materials
is equivalent irrespective of lighting. Thus, cyclists who
habitually wear fluorescent – as opposed to reflective – materials
may considerably overestimate their visibility at night. This may
result in cyclists unintentionally placing themselves at elevated
risk. Future research should explore the interaction between
bicycle lights and clothing to ascertain whether there are
differential effects on cyclist visibility at night.
Overall, the older drivers recognised cyclists significantly less
often than did younger drivers. While the younger drivers saw
the cyclists 51% of the time, older drivers identified them only
27% of the time and never identified them wearing black or
fluorescent clothing. The reduction in the ability of older
drivers to recognise the cyclists is likely to be due partly to
changes in visual function, especially age-related changes in
visual acuity and contrast sensitivity (ability to see faint
images), which may be exacerbated under low luminance
conditions [17].
Importantly, when the cyclists were wearing the vest with
reflectors on the ankles and knees, the older drivers recognised
them (80%) almost as often as did the younger drivers (100%).
The finding that cyclists are rarely seen by older drivers when
they are not wearing reflective clothing at night is important,
given the growing numbers of older drivers on our road
systems and the fact that many drive at night-time.
Collectively, the findings of this study provide important
preliminary data to suggest that cyclist visibility in low light is
strongly influenced by the clothing worn by the cyclist, and
highlight the importance of education among the general
population with regard to the utility of high-visibility clothing.
The data also underscore the fact that even alerted drivers
commonly fail to recognise the presence of cyclists, dependent
on the clothing configurations worn. These data also provide
evidence to support our previous findings with regard to
misunderstandings that cyclists have with regard to their own
visibility at night and suggest that cyclists may need to be better
informed with regard to the limits, as well as the benefits, of
specific visibility aids.
Acknowledgments
This research was funded by an Australian Research Council
Linkage Grant. The authors would like to express appreciation
to Queensland Transport for allowing the use of the facilities at
the Mt Cotton Driver Training Centre and to the staff of the
Mt Cotton Centre for their generous cooperation and support.
References
1. Joshi MS, Senior V, Smith GP. A diary study of the risk perceptions of
road users. Health, Risk & Society, 2001; 3: 261-79.
2. Watson L, Cameron M. Bicycle and motor vehicle crash
characteristics. Melbourne: Monash University Accident Research
Centre, 2006.
3. Sonkin B, Edwards P, Roberts I, Green J. Walking, cycling and
transport safety: An analysis of child road deaths. Journal of the Royal
Society of Medicine, 2006; 99: 402-05.
4. Jaermark S, Gregersen NP, Linderoth B.The use of bicycle lights,TFB
& VTI Forskning/Research, 1991.
5. Rodgers GB. Bicyclist deaths and fatality risk patterns. Accident
Analysis & Prevention, 1995; 27(2): 215-23.
6. Owens DA, Sivak M. Differentiation of visibility and alcohol as
contributors to twilight road fatalities. Human Factors, 1996; 38(4):
680-89.
7. Gale AG, Cairney PT.The conspicuity of bicycle lighting under on-road
conditions. In Vision in vehicles. Oxford: Elsevier Science Ltd, 1998.
8. Blomberg RD, Hale A, Preusser DF. Experimental evaluation of
alternative conspicuity-enhancement techniques for pedestrians and
bicyclists. Journal of Safety Research, 1986; 17: 1-12.
9. Räsänen M, Summala H. Attention and expectation problems in
bicycle-car collisions: An in-depth study. Accident Analysis &
Prevention, 1998; 30: 657-66.
10. Wood JM et al. Drivers' and cyclists' experiences of sharing the road:
Incidents, attitudes and perceptions of visibility. Accident Analysis &
Prevention, 2009; 41(4): 772-76.
11. Tyrrell RA, Wood JM, Carberry TP. On-road measures of pedestrians'
estimates of their own nighttime conspicuity. Journal of Safety
Research, 2004; 35(5): 483-90.
12. Johansson G. Visual perception of biological motion and a model for
its analysis. Perception & Psychophysics, 1973; 14(2): 201-11.
13. Wood JM, Tyrrell RA, CarberryTP. Limitations in drivers' ability to
recognize pedestrians at night. Human Factors, 2005; 47(3): 644-53.
14. Tyrrell RA et al. Seeing pedestrians at night: Visual clutter does not
mask biological motion. Accident Analysis & Prevention, 2009; 41(3):
506-12.
15. Balk SA et al. Highlighting human form and motion information
enhances the conspicuity of pedestrians at night. Perception, 2008;
37: 1276-84.
16. Jones K, Bentley BE, Wood JM, Woolf MI. Application of parallax for
the measurement of visibility distances in the open-road
environment. International Archives of Photogrammetry and Remote
Sensing, 1998; 32: 74-79.
17. Haegerstrom-Portnoy G, Schneck ME, Brabyn JA. Seeing into old
age:Vision function beyond acuity. Optometry & Vision Science,
1999; 76(3): 141-58.
ACRS Journal 21_No3:ACRS Journal Vol 17 No 2 10/8/10 2:06 PM Page 60
