Developmental Aspects of Children's Behavior and Safety While Cycling

Abstract

To examine children's competence while cycling, as demonstrated in mistakes in performance and failure to comply with safety rules.
Children in three age groups (8, 10, and 12 years) participated in a realistic yet simulated traffic environment.
The boys' cycling speed increased steadily with age, while that of the girls increased from 8 to 10 but decreased at age 12. Most children had adequate motor control by age 10, and the youngest compensated for their less developed skills by cycling slowly and braking early at junctions. Serious mistakes, often related to the children's age and gender, consisted of the children failing to stop at signals or stopping too late, especially at short stopping range.
There are considerable individual differences in children's cycling competence that are related to biological factors, such as age and gender, and psychological factors, such as rule compliance and choice of cycling speed.

Full text

Developmental Aspects of Children’s Behavior and
Safety While Cycling
Valdimar Briem, PHD, Karl Radeborg, PHD, Ilkka Salo, PHD, and Hans Bengtsson, PHD
Department of Psychology, University of Lund, Sweden
Objective To examine children’s competence while cycling, as demonstrated in mistakes in
performance and failure to comply with safety rules.
Methods Children in three age groups
(8, 10, and 12 years) participated in a realistic yet simulated traffic environment.
Results The boys’ cycling speed increased steadily with age, while that of the girls increased
from 8 to 10 but decreased at age 12. Most children had adequate motor control by age 10, and
the youngest compensated for their less developed skills by cycling slowly and braking early at
junctions. Serious mistakes, often related to the children’s age and gender, consisted of the
children failing to stop at signals or stopping too late, especially at short stopping range.
Conclusions There are considerable individual differences in children’s cycling competence
that are related to biological factors, such as age and gender, and psychological factors, such as
rule compliance and choice of cycling speed.
Key words children; cycling behavior; attention; safety rules; risk taking;
traffic environment.
Cycling accidents are among the most common causes of
physical injury to children. The accident rate is low
among younger children, who are generally not allowed
into traffic unsupervised, but increases as children grow
older and begin to cycle more on their own. Statistical
sources indicate that the cycling accident rate among
unprotected Swedish children comprises up to 80% of all
their traffic accidents, the rate steadily rising until the
age of about 12 or 13, and then falling off slowly. These
statistics also show that in the age range of 7–14, boys
have considerably more serious cycling accidents than
girls, the ratio being about 2:1 (Briem, 2003a). One
explanation for this difference is that boys normally have
a higher activity level and greater exposure to danger
than girls, and therefore more opportunities to have
serious accidents (Briem, 1988; Hargreaves & Davies,
1996). However, the rate of major and minor accidents is
liable to be shaped not only by activity and exposure, but
also by psychological factors associated with children’s
cognitive development, attitudes to risk, and under-
standing of task requirements (Briem, 2003b; Briem &
Bengtsson, 2000).
Cycling can be viewed as a skill consisting of several
different components, including motor elements such as
pedaling, balancing, steering, and braking, and cognitive
elements such as concentration, attention, judgment,
planning, and decision making. Even in situations in
which no extraneous distracting factors are involved,
bicycle riding requires the coordination of various
processes. Considerable alertness is required for the
successful selection of task-relevant information in
complex traffic situations, while ignoring distractions
(Rogers, Rousseau, & Fisk, 1999; Styles, 1997). Several
researchers (Eilert-Petersson & Schelp, 1997; Pless,
Taylor, & Arsenault, 1995) have argued that limitations
in attention capacity play a key role in causing traffic
accidents among young cyclists.
Modes of task performance differ. An individual
performing a novel task does so with conscious and
deliberate control. This requires a sizable proportion of
All correspondence should be sent to Valdimar Briem, Department of Psychology, University of Lund, Box 213,
S-221 00 Lund, Sweden. E-mail: Valdimar.Briem@psychology.lu.se.
Journal of Pediatric Psychology 29(5) pp. 369–377, 2004
DOI: 10.1093/jpepsy/jsh040
Journal of Pediatric Psychology vol. 29 no. 5 Q Society of Pediatric Psychology 2004 ; all rights reserved.
by guest on June 8, 2013http://jpepsy.oxfordjournals.org/Downloaded from

the individual’s limited cognitive processing capacity but
is also characterized by flexibility and adaptability. More
automatic control is found in tasks that have been
practiced over longer periods of time. Here, performance
tends to be quick and efficient, requiring little cognitive
processing, but is also fairly inflexible and less amenable
to conscious modification (Schneider & Shiffrin, 1977;
Shiffrin & Schneider, 1977). The former mode is likely
to be characteristic of the performance of the young,
novice cyclist, while the latter mode will be more likely
to be characteristic of older children with more cycling
experience.
Younger children can also be expected to have
difficulty sustaining a state of concentration in many
tasks, and to be susceptible to distraction by irrelevant
stimuli (Plude, Enns, & Brodeur, 1994; Sandels, 1970).
In the young cyclist, this may be seen in a difficulty in
attending simultaneously to more than one aspect of
a dangerous situation, such as when two cars approach
the point at which the child is about to cross the road
(Briem & Bengtsson, 2000).
There are various sources of distraction. Listening to
music (e.g., through a Walkman) while cycling can,
a priori, be expected to distract some children in this
complicated task. Previous studies have provided
somewhat conflicting evidence, some finding music to
improve and others to impair performance on tasks
requiring concentration (Beh & Hirst, 1999; Crawford
& Strapp, 1994; Daoussis & McKelvie, 1986; Fogelson,
1973; Mayfield & Moss, 1989; Schreiber, 1988; Wolf &
Weiner, 1972). Some have shown interaction effects
between the presence of music, the difficulty of the task,
and the child’s age (Higgins & Turnure, 1984). Two
effects stand out here, viz., that in some cases music
seems to cause cognitive overload, while in others the
sound of music stimulates the individual to mobilize
extra effort. Either way, music would seem to be
a suitable distractor in children’s cycling performance.
The way children manage cycling speed is likely to
be an important factor in their performance. This applies
both to the general choice of speed and the ability to
adapt one’s chosen speed to prevailing circumstances.
While increased speed brings about a reduction in the
time available for decisions and maneuvers, some
children will cycle faster than their capacity allows and
therefore be more exposed to accident risk.
One would expect cycling performance to improve
with age and increased cognitive and motor capacity,
taken together. Yet the accident statistics do not support
this observation (Briem, 2003a; Peterson, Gillies, Cook,
& Schick, 1994). This may suggest that older children
subject their more developed capacities to greater
demands than younger children, by cycling faster and/
or under more distracting conditions, thereby increasing
their accident risk (Peterson, Oliver, Brazeal, & Bull,
1995; Wilde, 1994).
The present study examines children’s cycling
behavior, particularly the capacity to react adequately
to both variable and systematic changes in attentional
requirements. Age- and gender-related differences were
hypothesized, both in the children’s ability to control the
bicycle and in the way situational demands and
distraction affect their behavior in critical situations.
Thus, older children were expected to have greater
confidence in their abilities, to take greater risks, and, as
a consequence, to sometimes make serious mistakes.
Girls were expected to be more careful than boys, to
cycle more slowly, and to make fewer serious mistakes.
Methods
Participants
A total of 57 children took part in the study, 26 girls and
31 boys, recruited from the second, fourth, and sixth
school years (mean ages 5 8.7, 10.6, and 13.1 years,
respectively) at a primary school in a town in southern
Sweden. Participation was voluntary and subject to the
approval of the school authorities and teachers and
written permission from the parents.
The Test Site and Materials Used
The study was conducted at a ‘‘traffic playschool,’’ which
is a test track that contains reduced-scale, yet realistic,
roads, bicycle lanes, pedestrian lanes, and an abundance
of road signs. It is located on the edge of a park near the
children’s school and is occasionally used for teaching
purposes. The layout is shown in Figure 1, and the basic
landmarks used in the experiment are marked on the
map. The track consists of a simple system of asphalt
roads, each divided into two traffic lanes, 2.2 m wide,
with a broken line down the middle. Two ‘‘rounds,’’ or
laps around the track, were employed: a whole round of
the outside of the track (b) and a half round through the
middle (c), their total lengths measuring 120 and 90 m,
respectively. At the middle of the track, four houses were
located around a four-way crossing, and these, along
with the trees and bushes, prevented the cyclists from
observing the entire track at any one time. Two
stationary signaling devices of standard design were
installed, one at the railway crossing (triangular sign with
370 Briem, Radeborg, Salo, and Bengsston
by guest on June 8, 2013http://jpepsy.oxfordjournals.org/Downloaded from

two red, flashing lights and a sound signal from a bell
[g]), the other at the four-way crossing (standard traffic
lights, red-amber-green [g]). There was fairly constant
traffic noise from nearby streets. The children’s behavior
was video-recorded from two positions (h), diagonally
from the front at both crossings (see Figure 1).
Many of the older children brought their own
bicycles to the test, while the younger children normally
used good, standard bicycles provided at the site,
individually adjusted to suit each child. These bicycles
all had pedal brakes, while some of the older children
had bicycles with hand-operated brakes. All the children
cycled the track wearing protective helmets.
Procedure
Before setting off, the children were instructed in groups
of two or three to cycle around the track, alternately
completing whole rounds (passing the railway crossing)
and half rounds (passing the four-way crossing). The
children were instructed to behave as if they were in
normal traffic, to pay particular attention to the traffic
lights, and to stop when the lights changed to red. They
were told to cycle at the speed they normally would
under comparable circumstances, for instance when
cycling to meet a friend. After the instructions, one of the
experimenters cycled ahead of the children for two
rounds, with signal changes on each round. The
instructions were concluded with the children cycling
alone twice, with the experimenter looking on.
During the experiment, the children took turns
cycling individually, and the children not cycling were
not allowed to watch, but sat inside one of the houses.
Each child took two turns at cycling (Parts 1 and 2), once
with and once without potential distraction in the form
of popular music played on a Walkman (using ear-
phones), the children themselves adjusting the volume
to a level they found comfortable. In Part 1, the child
cycled 11 rounds, alternating half and whole rounds,
beginning and ending with a whole round. Signals were
changed to red in 4 rounds (R): R5 (whole), R6 (half), R8
(half), and R9 (whole). Then the first child took a 7–10
minute break, sitting inside the house, while the second
child cycled his/her rounds in Part 1. The first child
returned to take his/her second turn (Part 2), cycling
another 11 rounds with four signal changes as before,
after which the second child concluded his/her Part 2.
The critical events recorded on video were the
children’s reactions when the signal changed to red at
the crossings. The two stopping distances before the stop
line in front of the signal lights are marked in Figure 1 as
e (short) and f (long) (see Independent Measures
below). The signal change was activated manually by
one of the experimenters when the children’s front
bicycle wheel reached the (discreetly) marked stopping
distances.
Figure 1. Map of the traffic playschool
prepared by the Technical Office,
Municipality of Helsingborg. Key 5 (a)
starting position, (b) whole round, (c)
half round, (d) stop lines, (e) short
stopping range, (f) long stopping range,
(g) signals, and (h) camera positions.
Children’s Behavior and Safety While Cycling 371
by guest on June 8, 2013http://jpepsy.oxfordjournals.org/Downloaded from

Dependent Measures
Cycling Speed
Time per round (TPR) in seconds was measured with
a stopwatch in both Parts 1 and 2 on four whole rounds
(R1, R3, R7, and R11) without a signal change, a total of
eight times in both parts together: R1 (after starting from
a standstill), R3 (before the first signal change), R7 (after
two signal changes), and R11 (after four signal changes).
The mean speed per round is reported in meters per
second (mps) 5 120/TPR (1 mps 5 3.57 km per hour).
Behavior at Sign al Changes
The recorded measures of the children’s performance
were:
The number of times they failed to brake to a standstill
before the stop line when the signal changed. This
included (1) the times when the children completely
failed to stop at a signal (missed signals); (2) the times
when they responded to the signal change but were
unable to stop in time, or stopped so late that the cycle
partly or completely passed the stop line (overshooting),
recorded either as the number of times the children
overshot or as a degree of overshooting on a scale from
1to3(15 more than a quarter but less than half the
bicycle, 2 5 more than half but less than the whole
bicycle, and 3 5 the whole bicycle beyond the line); and
(3) a combined measure indicating the occurrence of
either missed signals or overshooting on a given
occasion.
Their competence when braking at a signal measured by
whether they jumped down from the bicycle instead of
braking in a more conventional way. This included
placing one or both feet on the ground, even while the
bicycle was still moving quite fast, then walking,
running, or hopping along, holding the bicycle by the
handlebars until able to stop.
Independent Measures
Between-subject variables were age group (school years
2, 4, and 6) and gender. Age (8.0–13.7 years) and cycling
speed, in mps, were used as correlates in several
analyses. Within-subject variables were crossing (rail-
way crossing, four-way crossing), stopping range (long
[3.5 m], short [1.5 m]), and distraction (music, no
music). Requirement for fast reaction was explicitly
varied by activating the signals at either short or long
stopping range, and demands were placed on the
children’s attention capacity by having them listen to
music while cycling. The order of levels of all within-
subject independent variables was carefully counter-
balanced between age groups and genders.
Results
The level of significance was set at 5%. Post hoc testing
was done for all main factor differences, the Games-
Howell (
GH
) for between-subject and Scheffe
´
’s F (
SF
) for
within-subject differences. Significant analysis of vari-
ance (ANOVA) results are shown in Table I.
Cycling Speed
A child’s cycling speed was estimated as the average of
R3, R7, and R11 (a total of six rounds in Parts 1 and 2;
overall M 5 3.63 mps, SD 5 0.58, range 5 2.50–5.54).
A repeated-measures ANOVA indicated significant main
effects of age group, gender, and their interaction (see
Table I, Speed; and Figure 2). A breakdown analysis in
one-way ANOVAs indicated that the cycling speed
increased for boys between Years 2, 4, and 6 (p
GH
,
.05 in all three comparisons) and for girls between Years
2 and 4 (p
GH
, .05), but not Year 6. A significant
difference between boys and girls was found only in Year
6(p
GH
, .05).
Missed Signals
Nearly half the children (27) failed to stop at signal
changes at least once out of the eight rounds, (see Table
II). A one-sample sign test (hypothesized value 5 0)
indicated p , .01. No correlation was found between
cycling speed and number of missed signals, r(55) 5
.06, ns.
ANOVA results (Table I, Missed Signals) indicate
that the majority of missed signals occurred among the
girls, at short stopping range and at the four-way
crossing (Figure 3). Direct observation of the video
recordings in slow motion revealed that the girls
sometimes looked up at the signal light at the four-way
crossing from a distance of several meters before it
changed to red, then looked down and missed seeing the
light change. The children also missed the signal
disproportionately often at short range, at the railway
crossing, while listening to music.
Overshooting
On average, the children overshot more than twice
during eight rounds with signal changes (Table II).
While less than one fifth (9) succeeded in coming to
a halt on all eight occasions before the bicycle had partly
or wholly overshot the stop line (on the scale of 1–3; see
372 Briem, Radeborg, Salo, and Bengsston
by guest on June 8, 2013http://jpepsy.oxfordjournals.org/Downloaded from

Behavior at Signal Changes above; hypothesized value 5
1, p , . 01), more than two fifths of the children (23)
failed at least once to stop before the whole bicycle had
passed the stop line (gross overshootings 5 3, see
Behavior at Signal Changes above; hypothesized value 5
0, p , .01). The degree of overshooting was significantly
correlated with cycling speed, r(55) 5 .32, p , .05.
ANOVA results (Table I, Overshooting) indicate
that the tendency to overshoot was greatest at short
stopping range for the Year 4 girls and the Year 6 boys,
with a steady increase in degree of overshooting from age
8to12(p
GH
, .05). Whereas for the girls, overshootings
occurred mostly at the four-way crossing, for the boys
they occurred at both crossings.
Missed Signals and Overshooting
As many as 52 out of the 57 participants (91%) either
missed a signal or overshot the stop line at least once at
the eight signal changes (Table II) (sign test, hypothe-
sized value 5 0, p , .01; hypothesized value 5 2, p ,
.05). If we include only serious mistakes (i.e., the child
missed the signal or grossly overshot), we see that 41 of
the children (72%) committed these mistakes at least
one time out of eight (Table II) (sign test, hypothesized
value 5 0, p , .01).
ANOVA results for the absolute occurrence of
serious mistakes (Table I, Missed Signals and Over-
shootings) indicate that such mistakes happen mostly at
short stopping range, where they tend to increase with
age (p
GH
, .05 between Years 2 and 6), and to a much
lesser extent at long range, where they decrease with age
(p
GH
, .05 between Years 2 and 6). This is shown in
Figure 4. Also, the girls made more serious mistakes at
short range at the four-way crossing than the boys (cf.
Figure 3).
Jumping Down from the Bicycle at Signal Changes
ANOVA results (Table I, Foot-on-Ground Stops) in-
dicate that this inefficient braking method was employed
mainly by the younger children (p
GH
, .05) at long
stopping range at the railway crossing. Jumping down
was also significantly correlated with cycling speed,
r(55) 5 .32, p , .05.
Sequential Effects
Speed Throughout the Experiment. Cycling speed was
similar in Parts 1 and 2, with means of 3.38, 3.69, 3.66,
and 3.59 mps in the four timed rounds. It was
Table I. Results from Five Repeated-Measures Analyses of
Variance (Five Dependent Variables), Showing the Effects of the
Five Independent Variables
a
and Their Interactions
Dependent Variable Effect df F Power
Speed
Age group 2, 51 14.84
***
1.00
Gender 1, 51 8.40
**
0.82
Age Group 3 Gender 2, 51 5.84
**
0.86
Missed signals
Crossing 1, 51 16.46
***
0.99
Range 1, 51 33.79
***
1.00
Crossing 3 Range 1, 51 5.18
*
0.60
Gender 3 Crossing 1, 51 8.95
**
0.85
Gender 3 Range 1, 51 5.12
*
0.65
Gender 3 Crossing 3 Range 1, 51 8.28
**
0.89
Crossing 3 Range 3 Music 1, 51 7.52
**
0.78
Overshootings (degree = 1–3)
Age group 2, 51 3.81
*
0.67
Range 1, 51 85.91
***
1.00
Age Group 3 Gender 2, 51 9.31
***
0.98
Age Group 3 Gender 3 Crossing 2, 51 6.98
**
0.92
Age Group 3 Range 2, 51 5.25
**
0.82
Age Group 3 Gender 3 Range 2, 51 5.74
**
0.86
Missed signals and gross overshootings
Crossing 1, 51 14.29
***
0.97
Range 1, 51 51.68
***
1.00
Age Group 3 Range 2, 51 3.94
*
0.68
Gender 3 Crossing 3 Range 1, 51 4.29
*
0.52
Foot-on-ground stops
Age group 2, 51 3.33
*
0.60
Crossing 1, 51 13.61
***
0.97
Crossing 3 Range 1, 51 5.10
*
0.59
a
Two between-subject, three within-subject.
Only statistically significant effects are indicated.
* p , .05; ** p , .01; *** p , .001.
Figure 2. Average cycling speed (mps) as a function of the children’s
school year and gender. Standard errors are indicated.
Children’s Behavior and Safety While Cycling 373
by guest on June 8, 2013http://jpepsy.oxfordjournals.org/Downloaded from

significantly lower in R1 than in R3, R7, and R11 (p
SF
,
.01 in all three comparisons) and decreased from R3 to
R11 (p
SF
, .05).
Speed Related to Distraction. No effects on cycling speed
were found, sequential or otherwise, specifically attrib-
utable to the incidental music.
Sequential Effects in Missed Signals and Overshoot-
ing. The children missed the signal less often in Part
2(M 5 .09) than Part 1 (M 5 .16, p
SF
, .05), and less
often in R9 (M 5 .04) than in R5, R6, and R8 (Ms 5 .10,
.18, and .18; p
SF
5 .01). No effects were significantly
related to overshooting.
Discussion
A basic assumption here is that children’s cycling
performance is a joint function of their cognitive and
motor capacities and the demands placed on these
capacities, and that increased task load leads to a de-
terioration in performance. Task-load variation was
achieved through systematic variation in stopping range
and distraction and was further regulated by the
children’s own cycling speed. As the children grew
older, we found an increase not only in cycling speed,
but also in the number of mistakes they made. The
majority of mistakes occurred at short stopping range,
which required faster reactions.
Contrary to our hypotheses, increasing the pre-
sumed cognitive task load even further (i.e., using music
as a distractor) did not result in more mistakes due to
facilitation or interference. Possible reasons for this are
either that music is generally not a distractor for children
or that children are able to ‘‘tune out’’ irrelevant and
nondemanding stimuli when performing other, more
important tasks. The latter explanation receives support
in analogous studies of older individuals (Briem &
Hedman, 1995).
The children’s cycling mistakes were divided into
two main categories: overshootings and missed signals.
In the case of overshootings, the child noticed the signal
change but did not manage to stop until the bicycle had,
partly or wholly, passed the stop line. Speed alone
cannot satisfactorily explain all these mistakes. Clearly,
the children preferred to stop with the front wheel as
close to the line as possible, although generally not
across it. This was seen in the video recordings, where
Table II. The Number (%) of Children Who Made a Specified Number of More or Less Serious Mistakes at the Stop Line on 8 Trials
01234567M Odds
MS 30 (53) 8 (14) 10 (18) 8 (14) 1 (2) 1.0 1:8
OS 9 (16) 16 (28) 15 (26) 8 (14) 3 (5) 4 (7) 2 (4) 2.0 1:4
GOS 34 (60) 18 (32) 4 (7) 1 (2) 0.6 1:15
MS þ OS 5 (9) 9 (16) 9 (16) 9 (16) 13 (23) 9 (16) 2 (4) 1 (2) 3.0 1:3
MS þ GOS 15 (26) 18 (32) 10 (18) 10 (18) 3 (5) 1 (2) 1.5 1:6
MS = missed signals; OS = overshooting; GOS = gross overshooting.
The last two columns show the mean number of mistakes per child and the odds that a cycling child will make a mistake.
Figure 3. Mean number of signals missed by boys and girls,
respectively, at short and long stopping ranges, at the railway and four-
way crossings. Standard errors are indicated.
Figure 4. Mean number of serious mistakes (missed signals and gross
overshootings) in relation to stopping range and age group. Standard
errors are indicated.
374 Briem, Radeborg, Salo, and Bengsston
by guest on June 8, 2013http://jpepsy.oxfordjournals.org/Downloaded from

the children, while waiting for the light to change back
to green, often moved the bicycle backward behind the
stop line after having overshot. Cycling at a fairly high
speed, aiming to stop as close to the line as possible, is
bound to sometimes result in overshootings. Telling
children to cycle at a normal speed and not to race
clearly did not have an effect.
More than half of the children completely failed to
stop at least once in eight signal changes, thus making
a mistake that in real traffic would be associated with
high risk of being run over. This included both the times
when the children clearly did not see the signal and the
times they noticed the signal too late and did not bother
to stop. From the video recordings, we saw that the latter
sometimes occurred: The children, while cycling past
either of the crossings, would occasionally brake, shout
out, grimace, or grin sheepishly, but, being beyond the
stop line, would carry on cycling.
The children missed signals more often at the four-
way crossing, which, unlike the railway crossing, had
only a light signal. The significance of the sound signal
was corroborated by the fact that the children more
frequently missed the signal at short range at the railway
crossing when listening to music. It is possible that the
children at times relied on the sound of the bell to tell
them to stop at the railway crossing, causing them to pay
less attention to the light or even disregard it altogether.
Obviously, listening to music while cycling may serve to
mask other, important auditory events, in a way that
compromises the children’s safety.
The boys generally made more speed-related mis-
takes than the girls. The girls, however, missed more
signals than the boys, primarily girls in Years 4 and 6 at
the four-way crossing. In the video recordings, we see
these girls cycling slowly round the track. At the four-
way crossing before the light changed, they would
sometimes quickly look up, then down again, cycling
straight past, apparently without any idea that the light
had changed to red, a remarkable lapse of attention. In
contrast, the boys, cycling considerably faster, often
appeared more vigilant, directing their gaze at the signal
as the light changed to red.
Real-life accidents happen under circumstances
similar to the ones observed here, as indicated in several
studies of child cyclists. Several investigators studying
the patterns of serious accidental injuries sustained by
child cyclists (Fife et al., 1983; Nixon, Clacher, Pearn, &
Corcoran, 1987; Williams, 1976) found that the most
common kind of car/bicycle collision occurred when
a child, suddenly and without warning, cycled into the
path of an oncoming automobile on the near side of the
road. Under these conditions, the car driver would in
many cases have had little chance of discovering the
cyclist until it was too late to avoid a collision. The
responsibility for the accident is then usually ascribed to
the cyclist. However, the accident risk in these
encounters is further aggravated by car drivers’ not
being fully aware of this kind of cyclist behavior. For this
reason, Summala, Pasanen, Raesaenen, and Sievaenen
(1996) suggested that drivers develop a visual scanning
strategy that concentrates on the detection of frequently
occurring major hazards at the expense of visual
information associated with less frequent risks. Colli-
sions with suddenly appearing cyclists would seem to fall
into the latter category.
Missed signals and overshootings are both potential
causes of accidents in real traffic. In the absence of other
information, one might presume that both are caused by
the same factors, but the results demonstrate that this is
not the case, as, for instance, only overshooting is
significantly correlated to cycling speed. There is some
indication that some children ‘‘favor’’ one kind of
mistake, but not both. A joint analysis of these serious
mistakes shows a clear increase with age at the shorter
stopping range, when the children are subjected to
a greater task load, but a decrease with age at the longer
range, where decision time is longer and the task load is
smaller.
Although the youngest children generally made
fewer mistakes, their cycling skills were fairly rudimen-
tary. For instance, stopping the bicycle by placing one or
both feet on the ground is normally a fairly inefficient
strategy, which mainly the younger children employed,
probably because they had not yet learned to make
efficient use of the bicycle brakes. When employed at
long range, especially at slow speed, this mode was fairly
successful, whereas at short range it often resulted in the
child overshooting. Observation of the video recordings
also showed that some of the youngest children carefully
anticipated the signal changes, reducing their speed or
braking at a distance of several meters, sometimes to
a standstill, when approaching a crossing.
A cyclist’s performance in a given traffic situation is
determined not only by his or her motor and perceptual
capabilities, as some researchers have implied (Drury &
Daniels, 1980). Motivational factors determine how
these abilities are utilized, and the need for attention
to the task at hand is often overshadowed by other
objectives, in ways that significantly affect children’s
accident proneness. It has been pointed out (Peterson,
Children’s Behavior and Safety While Cycling 375
by guest on June 8, 2013http://jpepsy.oxfordjournals.org/Downloaded from

Brazeal, Oliver, & Bull, 1997; Peterson et al., 1994) that
as children grow older, they gain confidence as cyclists,
at the same time as their fear of possible dangerous
consequences decreases. Diminished apprehension with
regard to the consequences of one’s actions, as well as
changes in attitudes toward the social desirability of
daring and spectacular actions, may lead to increased
risk taking. Consequently, older children do not
always perform better, even though their physical and
cognitive capacity is superior to that of younger
children.
The children in the present study were aware that
they would not come to any serious harm if they
happened to stretch the safety rules. This knowledge
may have contributed to making their risk acceptance
higher and lowered their vigilance. In real traffic, very
few such mistakes actually lead to accidents, probably
because there is no vehicle present at that particular
moment. Children’s beliefs in their own invulnerability
and actions resulting from such beliefs can have fatal
consequences. Most accidents actually happen in
environments where there is not much traffic, where
children feel safe, and where they feel that playful risk
taking does not matter.
The simulated and danger-free traffic environment
in which the children’s cycling behavior was studied
allowed good control of the experimental conditions and
observation. The results confirm and complement those
of earlier, comparable studies (Peterson et al., 1997;
Peterson et al., 1994; Peterson et al., 1995). They
demonstrate that children often ignore safety rules while
cycling and that they will make mistakes, some children
more than others, but can also learn to avoid these
mistakes. As has been shown in earlier research (van
Schagen & Brookhuis, 1994), learning from experience
may, in this context, be considerably more effective than
theoretical knowledge. However, all things considered,
parents and others responsible for children’s safety
should exercise extreme caution and think twice before
letting children cycle in dangerous traffic.
Acknowledgments
This study was funded by a grant from Vinnova,
Stockholm. The authors wish to thank all the children
and teachers at Magnus Stenbock School in Helsingborg
who assisted in the study. Thanks are also due to the
Municipality of Helsingborg for making the traffic
playschool available, and to the Technical Office for
helping to install the signaling equipment.
References
Beh, H. C., & Hirst, R. (1999). Performance on
driving-related tasks during music. Ergonomics, 42,
1087–1098.
Briem, V. (1988). Barn i trafiken: Aktivitetsmo
¨
nster och
sa
¨
kerhet pa
˚
va
¨
gen till skolan. [Children in traffic:
Activity patterns and safety on the journey to
school]. Bulletin 78. Lund, Sweden: Department of
Technology and Society, Traffic Engineering, Lund
Institute of Technology.
Briem, V. (2003a). Barns sa
¨
kerhet i trafiken. [Children’s
traffic safety]. Unpublished manuscript (final report
to Vinnova, the Swedish agency for innovation
systems).
Briem, V. (2003b). The role of psychobiological factors in
children’s accidents. Manuscript in preparation.
Briem, V., & Bengtsson, H. (2000). Cognition and
character traits as determinants of young
children’s behaviour in traffic situations.
International Journal of Behavioral Development,
24, 492–505.
Briem, V., & Hedman, L. R. (1995). Behavioural effects
of mobile telephone use during simulated driving.
Ergonomics, 38, 2536–2562.
Crawford, H. J., & Strapp, C. M. (1994). Effects of vocal
and instrumental music on visuospatial and verbal
performance as moderated by studying preference
and personality. Personality and Individual
Differences, 16, 237–245.
Daoussis, L., & McKelvie, S. J. (1986). Musical
preferences and effects of music on a reading
comprehension test for extraverts and
introverts. Perceptual and Motor Skills, 62,
283–289.
Drury, C. G., & Daniels, E. B. (1980). Predicting bicycle
riding performance under controlled conditions.
Journal of Safety Research, 12, 86–95.
Eilert-Petersson, E., & Schelp, L. (1997). An
epidemiological study of bicycle-related
injuries. Accident: Analysis and Preventi on, 29,
363–372.
Fife, D., Davis, J., Tate, L., Wells, J. K., Mohan, D., &
Williams, A. (1983). Fatal injuries to bicyclists: The
experience of Dade County, Florida. Journal of
Trauma, 23, 745–755.
Fogelson, S. (1973). Music as a distractor on
reading-test performance of eighth grade
students. Perceptual and Motor Skills, 36(3 Pt. 2),
1265–1266.
376 Briem, Radeborg, Salo, and Bengsston
by guest on June 8, 2013http://jpepsy.oxfordjournals.org/Downloaded from

Hargreaves, D. J., & Davies, G. M. (1996). The
development of risk-taking in children. Current
Psychology: Developm ental, Learning, Personality,
Social, 15, 14–29.
Higgins, A. T., & Turnure, J. E. (1984). Distractibility
and concentration of attention in children’s
development. Child Development, 55, 1799–1810.
Mayfield, C., & Moss, S. (1989). Effect of music tempo
on task performance. Psychological Reports,
65(3 Pt 2), 1283–1290.
Nixon, J., Clacher, R., Pearn, J., & Corcoran, A. (1987).
Bicycle accidents in childhood. British Medical
Journal (Clinical Research Edition), 294, 1267–1269.
Peterson, L., Brazeal, T., Oliver, K., & Bull, C. (1997).
Gender and developmental patterns of affect,
belief, and behavior in simulated injury events.
Journal of Applied Developmental Psychology, 18,
531–546.
Peterson, L., Gillies, R., Cook, S. C., & Schick, B. (1994).
Developmental patterns of expected consequences
for simulated bicycle injury events. Health
Psychology, 13, 218–223.
Peterson, L., Oliver, K. K., Brazeal, T. J., & Bull, C. A.
(1995). A developmental exploration of
expectations for and beliefs about preventing bicycle
collision injuries. Journal of Pediatric Psychology, 20,
13–22.
Pless, I. B., Taylor, H. G., & Arsenault, L. (1995).
The relationship between vigilance deficits and
traffic injuries involving children. Pediatrics, 95,
219–224.
Plude, D. J., Enns, J. T., & Brodeur, D. (1994). The
development of selective attention: A life-span
overview. Acta Psychologica, 86, 227–272.
Rogers, W. A., Rousseau, G. K., & Fisk, A. D. (1999).
Applications of attention research. In F. T. Durso
(Ed.), Handbook of applied cognition. Chichester,
UK: Wiley.
Sandels, S. (1970). Young children in traffic.
British Journal of Educational Psychology, 40,
111–116.
Schneider, W., & Shiffrin, R. M. (1977). Controlled and
automatic human information processing: I.
Detection, search, and attention. Psychological
Review, 84, 1–66.
Schreiber, E. H. (1988). Influence of music on college
students’ achievement. Perceptual and Motor Skills,
66, 338.
Shiffrin, R. M., & Schneider, W. (1977). Controlled
and automatic human information processing: II.
Perceptual learning, automatic attending and
a general theory. Psychological Review, 84, 127–190.
Styles, E. A. (1997). The psychology of attention. Hove,
UK: Psychology Press/Erlbaum.
Summala, H., Pasanen, E., Raesaenen, M., & Sievaenen,
J. (1996). Bicycle accidents and drivers’ visual
search at left and right turns. Accident Analysis and
Prevention, 28, 147–153.
van Schagen, I., & Brookhuis, K. A. (1994). Training
young cyclists to cope with dynamic traffic
situations. Accident Analysis and Prevention, 26,
223–230.
Wilde, G. J. S. (1994). Target risk: Dealing with
the danger of death, disease and damage in
everyday decisions. Toronto, Canada:
PDE Publications.
Williams, A. F. (1976). Factors in the initiation of
bicycle–motor vehicle collisions. American Journal of
Diseases of Children, 130, 370–377.
Wolf, R. H., & Weiner, F. F. (1972). Effects of four noise
conditions on arithmetic performance. Perceptual
and Motor Skills, 35, 928–930.
Children’s Behavior and Safety While Cycling 377
by guest on June 8, 2013http://jpepsy.oxfordjournals.org/Downloaded from