Article Text

Download PDFPDF

Roles of individual differences and traffic environment factors on children’s street-crossing behaviour in a VR environment
  1. Huarong Wang1,
  2. Zhan Gao2,
  3. Ting Shen2,
  4. Fei Li3,
  5. Jie Xu3,
  6. David C Schwebel4
  1. 1 Traffic Psychology,Institute of Special Environmental Medicine, Nantong University, Nantong, China
  2. 2 School of Information Science and Technology, Nantong University, Nantong, China
  3. 3 Xinlin College, Nantong University, Nantong, China
  4. 4 Psychology, University of Alabama at Birmingham, Birmingham, Alabama, USA
  1. Correspondence to Dr Huarong Wang, Traffic Psychology, Nantong University, Nantong 226019, China; yeluo801004{at}163.com

Abstract

Objective Pedestrian injuries are among the most common cause of death and serious injury to children. A range of risk factors, including individual differences and traffic environment factors, has been investigated as predictors of children’s pedestrian behaviours. There is little evidence examining how risk factors might interact with each other to influence children’s risk, however. The present study examined the independent and joint influences of individual differences (sex and sensation seeking) and traffic environment factors (vehicle speeds and inter-vehicle distances) on children’s pedestrian safety.

Methods A total of 300 children aged 10–13 years were recruited to complete a sensation-seeking scale, and 120 of those were selected for further evaluation based on having high or low sensation-seeking scores in each gender, with 30 children in each group. Children’s pedestrian crossing behaviours were evaluated in a virtual reality traffic environment.

Results Children low in sensation seeking missed more opportunities to cross and had longer start gaps to enter the roadway compared with those high in sensation seeking, and these effects were more substantial when vehicles were spread further apart but travelling slowly. Interaction effects between inter-vehicle distance and vehicle speed were also detected, with children engaging in riskier crossings when the car was moving more quickly and the vehicles were spread further than when the vehicles were moving quickly but were closer together. No sex differences or interactions emerged.

Conclusion Both sensation seeking and traffic environment factors impact children’s behaviour in traffic, and there are interactions between traffic speeds and inter-vehicle distances that impact crossing behaviour.

  • virtual reality environment
  • pedestrian crossing behaviour
  • individual difference
  • Traffic environment
View Full Text

Statistics from Altmetric.com

Introduction

Road traffic crashes are the leading cause of death and injury to children in China.1 Among road traffic injury causes, pedestrian injury is the most common cause of mortality for Chinese children. Global Burden of Disease 2016 data estimate 68.74% of Chinese children aged 5–14 years who die in road traffic injuries are pedestrians rather than other road users.2

Previous research has summarised a range of risk factors that influence children’s pedestrian injuries.3 Among these, the effects of individual differences in temperament and personality are commonly cited. For example, several studies indicate sensation seeking is linked to traffic crash involvement and traffic offences among children and adolescents.4–6

Child sex is also mentioned frequently as a risk factor for child pedestrian injury,7–10 although empirical data supporting this assertion is somewhat inconclusive. For example, Fu and Zou8 observed that boys were more likely to run while crossing the street than girls and Sullman et al found that boys played on the road more often than girls,10 but other studies report no sex differences in children’s pedestrian safety behaviours.11–13 Several factors might explain conflicting previous results (eg, different age ranges studied and different pedestrian tasks),14 including biological or evolutionary-driven sex differences in risk-taking or impulsivity.15 16

Behavioural theory suggests that risk behaviours are likely influenced by multilayered and multifactored rather than single discrete factors,17 18 but unpacking the multilayered and multifactored individual difference factors that predict children’s safety in pedestrian environments is complicated. One complication is that pedestrian behaviour is influenced not just by individual differences in sex and personality/temperament, but also by children’s cognitive development. For example, perception and judgement of relevant stimuli, such as vehicle speed and inter-vehicle distances, impact all children’s crossing behaviours.19–21

Research specifically comparing the influence of vehicle speeds and distances on children’s crossing behaviours tends to suggest that vehicle distance is used more often than speed by children judging traffic safety. For example, Simpson et al studied pedestrian behaviour among children (5–9 and 10–14 years old), adolescents (15–19 years old) and adults (above 19 years old) in a virtual traffic environment. Vehicles travelled either at a uniform distance apart or at a uniform speed. The study results suggested that pedestrians of all ages based road crossing decisions on distance more than vehicle speed.19 Others report similar findings,20 22 although Velde et al suggested that children rely both on vehicle distance and speed to judge pedestrian safety.23 24

Given the apparent independent and joint influences of both individual differences and cognitive judgement of vehicle speed and distances on children’s pedestrian safety, we designed a study to evaluate both the independent and interaction effects among them. To test our hypotheses, we used a virtual pedestrian environment that allowed us to (1) evaluate behaviour safety without the actual risk of injury and (2) manipulate traffic speed and inter-vehicle distances in a controlled manner. We posited three primary hypotheses: (1) Boys and children with high sensation-seeking levels will take more risks crossing a virtual road; (2) Both inter-vehicle distance and vehicle speed affect children’s road-crossing behaviours; and (3) Given the lack of previous evidence, we posed an exploratory hypothesis that individual difference and vehicle speed/distance factors would interact to influence children’s crossing behaviours.

Methods

Participants

A total of 300 children in fourth and fifth grade were recruited from an elementary school in Nantong city, China. To select children with varying levels of sensation seeking, we administered the Chinese version of the Primary and Middle School Students’ Sensation Seeking Scale to all fourth and fifth grade students in the school (167 boys and 133 girls) and then selected 120 students to participate in the remainder of the study based on their scores.25 In total, 30 boys and 30 girls scoring high in sensation seeking (top 17.96% and 22.56% for boys and girls, respectively) and 30 boys and 30 girls scoring low in sensation seeking (bottom 17.96% and 22.56% for boys and girls, respectively) were included.

The average sensation-seeking score for boys in the low sensation-seeking group was 28.63 (SD=1.94, range=27–36), and for girls in the low sensation-seeking group was 29.42 (SD=2.43, range=27–36). The average score in the high sensation-seeking group for boys was 62.87 (SD=9.89, range=50–97), and for girls was 60.97 (SD=5.92, range=53–73). As expected, multivariate analysis of variance found that children’s scores between the high sensation-seeking and low sensation-seeking groups were statistically significant (F(3,116)=304.25, p<0.001, η2 p =0.89), and there was no significant difference between boys and girls in sensation seeking (F(1,116)<1) . There also was no statistically significant interaction effect between sex and group (F(1,116)=2.07, p>0.05).

The research design created four groups, therefore: boys high in sensation seeking, boys low in sensation seeking, girls high in sensation seeking and girls low in sensation seeking. There were 30 children in each group. The sample had a mean age of 11.7 years (SD=0.82 years; range=10–13 years old).

Written parental consent and verbal child assent were secured prior to each child’s participation. Children were compensated with small gifts, such as school supplies.

Virtual environment

An urban virtual traffic environment was created with buildings, roads and vehicles. The virtual environment was viewed through HTC VIVE, a virtual reality (VR) display which contains dual Active-Matrix Organic Light-Emitting Diode (AMOLED) screens with diagonal diameters of 3.6 inches. The head-mounted resolution of the 90 Hz head-mounted display is 1080×1200 pixels (combined resolution is 2160×1200 pixels). Its field of view is 110°. Sensors for the helmet displays include Steam VR tracking technology, G-sensor calibration, gyroscopes and proximity distance sensors. The system includes two VIVE locators that use 360° positioning coverage and wireless synchronisation technology.

The virtual environment consisted of a straight, flat section of roadway, several intersections that could be used for experimentation, and typical local buildings, streetlights and vehicles. The road itself was 5.6 m wide, presenting bidirectional traffic on two lanes. The lanes were marked with continuous white edge lines, dashed central white stripes and zebra crossings at intersections. Vehicles included typical models of private cars and taxies, all of which were 1.6 m in width and 3 m in length. Figure 1 illustrates a car crossing the zebra crossing. The sound of traffic in the environment corresponded to vehicle distances and speeds, offering Doppler sound effects to increase realism. When a child was struck by a vehicle, the screen instantly became grey and the sound of an ambulance siren was heard.

Figure 1

Schematic representation of virtual reality traffic scene.

Four types of traffic flow were presented to children, varying by inter-vehicle distances and vehicle velocities that represented common traffic behaviour on urban Chinese roads.

  1. Vehicles driving at 5 m/s and inter-vehicle distance was 75 m.

  2. Vehicles driving at 5 m/s and inter-vehicle distance was 130 m.

  3. Vehicles driving at 10 m/s and inter-vehicle distance was 75 m.

  4. Vehicles driving at 10 m/s and inter-vehicle distance was 130 m.

In all traffic flow conditions, the vehicles drove at constant and equivalent speeds with consistent inter-vehicle distances in both directions.

Experimental design

We implemented a 2 (sex: female vs male)×2 (sensation seeking: high vs low)×2 (inter-vehicle distance: 75 vs 130 m)×2 (vehicle speed: 5 vs 10 m/s) mixed factorial design. Sensation seeking and sex served as between-subject predictors, and vehicle speed and inter-vehicle distance as within-subject predictors.

Procedure

Each child was tested individually by a group of three trained research assistants. One research assistant worked directly with the child on the VR street-crossing trials. That assistant’s roles included guiding children on how to wear the VR headset, informing them of the task requirements and monitoring the safety of children during the experimental process. The second research assistant oversaw the operation of the computer that controlled the VR equipment and the third research assistant used a high-definition video camera to film children’s behaviour during the experimental sessions.

Following consent procedures, each session began by fitting children with the VR head-mounted display headset. Children were told that they should try their best to cross the road safely to the other side. Instructions were given verbally and children were offered the opportunity to ask questions. Children were then given two practice trials, both of them at intersections similar to those used in the experimental trials and with vehicles travelling at 8 m/s and 100 m apart from each other. A research assistant instructed children to wait on the curb and observe the traffic, then to start crossing the road when they felt it was safe. Once the child crossed to the other side of the road, the research assistant guided the child back to the curb and initiated the next practice trial. If a child was struck by a vehicle, the research assistant assured the child it was fine since this was a simulation, guided him/her to safety and then gave the child another practice trial.

Following the practice trials, children completed four sets of three unique trials following a procedure similar to that used in the practice trials. In total, therefore, each child completed 12 experimental trials (3 crossings by 4 traffic flows). The order of traffic flows was randomly assigned. During all experimental trials, data concerning children’s decisions were recorded automatically by the VR computer system, offering rich data on pedestrian behaviour and decisions. Specific outcome measures are described below.

On average, experimental sessions lasted approximately 30 min. The full experimental process was videotaped, enabling subsequent behavioural coding of children’s behaviour. Videotape coding was performed by two trained research assistants, who established inter-rater reliability through independent viewing of the full sample (r=0.98 for the variable ‘running’ and r=0.94 for the variable ‘missed opportunity’). Rare disagreements were resolved through discussion between the coders and a third senior researcher.

Measures

Sensation seeking

The Primary and Middle School Students’ Sensation Seeking Scale was used to measure children’s sensation-seeking level, which has documented reliability and validity data.25 The scale includes two dimensions: Thrill and Adventure Seeking, which reflects a desire to engage in sports or other activities involving speed or danger (eg, parachute jumping and surfing), and Disinhibition, which reflects desire to engage in antisocial negative risk-taking behaviours (eg, engage in illegal activities and behave rudely to teachers and elders). The scale includes 30 items, each scored on a 3-point scale (0, 1 and 2) with higher scores indicating higher levels of sensation seeking. The scale had strong internal consistency (α=0.91) in our sample.

Crossing behaviours

The data recorded by the computer and those coded from videotapes were used to calculate the following pedestrian safety outcome measures for each child, with results averaged across the three crossings in each traffic flow condition:

  1. Dangerous crossings (count)—number of crossings when the child was either struck by a virtual car or had a close call, which was defined as being within 1 s of being struck by an approaching vehicle.26 27

  2. Start gap (seconds)—for the traffic gap that the child chose to enter, the time elapsed between the last vehicle leaving the crosswalk until the child’s foot contacted the road from off the curb. This temporal lag before initiation of crossing into a traffic gap is considered an indicator of children’s efficiency of cognitive processing in pedestrian situations (also called start delay in the literature).27 28

  3. Running (count)—number of crossings during which the child ran while crossing the road, defined as both feet being off the ground simultaneously. This kind of behaviour is relevant to safety for two primary reasons: (a) running can lead to failure to observe oncoming traffic29 and (b) running increases risk of falling while crossing. Running children may also confuse drivers, leading to vehicle–child pedestrian crashes because drivers respond incorrectly.

  4. Missed opportunities (count)—number of times the child chose not to cross within a traffic gap that was 1.5 times or greater than the time required for the child to cross the street safely.30 31

Data analysis

Before evaluating the influence of multiple factors on children’s crossing behaviour, data cleaning excluded 19 children due to a technical or experimental error with the VR system. Analyses comparing the 101 valid children for analysis to the 19 excluded children found that they were not statistically different in terms of age, sex or sensation-seeking scores.

Primary analyses were conducted through a series of repeated measures analysis of variance (ANOVA) models with sex, sensation-seeking level, inter-vehicle distance and vehicle speed serving as the independent variables and the four pedestrian behaviour outcomes as the dependent variables. Both main effects and interactions between multiple factors were tested.

Results

A total of 101 children, including 27 boys high in sensation seeking, 27 boys low in sensation seeking, 25 girls high in sensation seeking and 22 girls low in sensation seeking, were included in the primary analyses. Each child completed 12 trials, 3 at each of the 4 traffic flows.

Dangerous crossing

We considered dangerous crossing first. A total of 82 of the 1212 valid trials resulted in collisions (6.8%) and 136 trials involved close calls (child was within 1 s of being struck; 11.2%). We combined the numbers of collisions and close calls and then averaged across trials for each of the four types of traffic flow to yield a measure of the percentage of dangerous crossings for each child in each traffic flow. Descriptive data appear in table 1.

Table 1

Descriptive statistics: mean (SE) for dangerous crossing (percentage) as a function of traffic flow types, sex and sensation-seeking level group

A 2 (sex: female vs male)×2 (sensation seeking: high vs low)×2 (inter-vehicle distance: 75 vs 130 m)×2 (vehicle speed: 5 vs 10 m/s) repeated measures ANOVA on the inter-vehicle distance and vehicle speed factors was computed. The percentage of dangerous crossings served as the dependent variable.

The results indicated statistically significant effects of inter-vehicle distance (Fdistance(1,97)=10.99, p=0.001, η2 p =0.10) and vehicle speed (Fspeed(1,97)=110.97, p=0.000, η2 p=0.53) on children’s dangerous crossings. We also detected a statistically significant interaction effect between the two vehicle variables, Fdistance×speed(1,97)=11.13, p=0.001, η2 p =0.10. Simple effect tests showed that when the car was moving more quickly and the vehicles were spread further, children engaged in riskier crossings than when the vehicles were moving quickly but were closer together (M=20% vs 11% for 10 m/s; p<0.05). When the car was moving more slowly, children engaged in dangerous crossings at similar rates (M=2% vs 2% for further vs closer distances at 5 m/s; p>0.05) (figure 2). There were no statistically significant effects of sex or sensation-seeking level, Fs(1,97)<1.

Figure 2

Percentage of dangerous crossings as a function of inter-vehicle distance and vehicle speed.

Start gap

A 2 (sex: female vs male)×2 (sensation seeking: high vs low)×2 (inter-vehicle distance: 75 vs 130 m)×2 (vehicle speed: 5 vs 10 m/s) repeated measures ANOVA analysed the effects of the factors of interest on start gap, or children’s temporal delay to enter the selected traffic gap after it appeared. Descriptive data appear in table 2.

Table 2

Descriptive statistics: mean (SE) for start gap (seconds) as a function of traffic flow types, sex and sensation-seeking level group

The results showed that sensation-seeking level had a pronounced effect on the ‘start gap’ measure: Children scoring low in sensation seeking waited much longer to enter gaps than those scoring high in sensation seeking (M=9.27 vs 8.39 s), F(1,97)=5.35, p=0.023, η2 p =0.06. Moreover, statistically significant differences in start gap also emerged based on inter-vehicle distance and vehicle speed, Fdistance(1,97)=27.33, p=0.001, η2 p=0.22 and Fspeed(1,97)=6.83, p=0.01, η2 p =0.07, and there was a statistically significant interaction effect between inter-vehicle distance and vehicle speed, Fdistance×speed(1,97)=59.93, p=0.000, η2 p =0.41. Simple effect analysis showed that children tended to have shorter delays starting into a traffic gap when vehicles were spaced closer together (M=6.36 vs 13.21 s) and when vehicles were moving more quickly (M=6.79 vs 12.77 s).

Finally, a statistically significant three-way interaction between inter-vehicle distance, vehicle speed and sensation seeking emerged, Fdistance×speed×sensation-seeking level(1,97)=8.84, p=0.004, η2 p =0.09. As shown in figure 3, sensation seeking was most relevant in the ‘easiest’ condition when vehicles were spread apart and travelling slowly. In the 130 m inter-vehicle distance and 5 m/s condition, there was a statistically significant difference between high (M=13.02 s) and low (M=16.53 s) sensation-seeking levels, F(1,97)=8.10, p=0.005, which was not statistically significant in the other traffic conditions (p>0.05). There, also, were no statistically significant sex differences, Fs(1,97)<1.

Figure 3

Mean seconds of start gap as a function of sensation-seeking level, inter-vehicle distance and vehicle speed.

Running

Children ran across the road on 154 of the 1212 valid trials (12.7%). A 2 (sex: female vs male)×2 (sensation seeking: high vs low)×2 (inter-vehicle distance: 75 vs 130 m)×2 (vehicle speed: 5 vs 10 m/s) repeated measures ANOVA yielded just one statistically significant effect, for vehicle speed, F(1,97)=8.44, p=0.004, η2 p =0.07. Post hoc Bonferroni tests indicated children ran more often when they faced vehicles travelling more quickly (10 m/s) than when they faced vehicles travelling more slowly (5 m/s; M=14.7% vs 11.3%). No other main effect was statistically significant, Fs(1,97)<1.

Missed opportunities

Children missed at least one safe opportunity to cross the street on 325 of 1212 trials (26.82%), including 24 trials in which children missed 10 or more safe opportunities to cross. We computed the average number of missed opportunities to cross within each traffic flow condition for analysis.

As presented in table 3, children with low sensation seeking (M=1.16 missed opportunities per crossing trial, SE=0.12) missed more opportunities than those with high sensation-seeking level (M=0.58 missed opportunities per crossing trial, SE=0.11), F(1,97)=7.29, p=0.008, η2 p =0.08. There were also statistically significant main effects of inter-vehicle distance and vehicle speed, Fdistance(1,97)=27.33, p=0.001, η2 p =0.22 and Fspeed(1,97)=6.83, p=0.01, η2 p =0.07. Moreover, there was a statistically significant interaction effect between inter-vehicle distance and vehicle speed, Fdistance×speed(1,97)=13.65, p=0.000, η2 p =0.11. Simple effect analysis showed that children tended to miss more opportunities to cross when vehicles were spaced close together and travelling quickly (p<0.001). No statistically significant sex differences were present (F(1,97)<1), nor did any other interaction effects emerge among the factors (p>0.05).

Table 3

Descriptive statistics: mean (SE) for missed opportunities as a function of traffic flow type, sex and sensation-seeking level group

Discussion

We investigated the direct and interacting roles of fourth-grade and fifth-grade Chinese children’s individual differences (sex and sensation seeking) and the traffic environment (traffic speeds and inter-vehicle distance) on children’s behaviour while crossing a virtual street. The findings reinforce the notion that paediatric pedestrian behaviour is not a single activity based on single risk factors, but rather represents a complex interplay of individual differences and the traffic environment that together impact children’s safety.

The effect of individual factors on pedestrian behaviours

Similar to previous findings,5 32 our findings confirm that sensation seeking plays a role in children’s pedestrian behaviours. Children low in sensation seeking missed more safe opportunities to cross, and they tended to hesitate much longer prior to starting to cross at traffic gaps compared with those high in sensation seeking. However, unlike previous research, our results did not find that sensation seeking was related to running or dangerous crossing. In other words, our results suggest that children high in sensation seeking may have been more efficient in their crossing—they entered gaps more quickly and missed fewer safe opportunities to cross—but they did not actually experience higher risk of injury or crash compared with children lower in sensation seeking. Children lower in sensation seeking were more cautious, leading to diminished efficiency, but were equally safe in their pedestrian decisions.

The impact of sensation seeking on children’s behaviour also varied somewhat across traffic flow environments. Specifically, sensation seeking most impacted pedestrian behaviour when vehicles were travelling slowly and at greater distances apart. In that traffic flow condition—the condition that might be described as ‘easiest’ or ‘safest’—children low in sensation seeking had much longer delays before starting their crossing and left a shorter gap before traffic compared with children with high sensation-seeking levels. Thus, sensation seeking influenced behaviour when cautious behaviour still permitted safety. When the traffic environment presented was more challenging to cross within, even those with lower sensation-seeking levels seemed to take some risk to maintain their safety. In those situations, individuals low in sensation seeking behaved more like their peers with higher sensation-seeking levels. Of course, these results must be interpreted within the context that the children were aged 10–13 years, and therefore should have developed the cognitive capacity to cross streets safely.11 24 33 Given their cognitive capacity to cross safely, sensation seeking influenced behaviour only when the children faced an environment that permitted greater caution within the context of safety. Results might differ among younger children whose cognitive skills to cross safely are less sophisticated.

Our results did not yield sex-based differences in children’s pedestrian behaviour. The existing literature is mixed on this topic, with some studies reporting boys take more risks in pedestrian environments than girls7 8 10 30 but others reporting no sex differences.12 13 One aspect of our study was unique compared with previous work. We purposively selected both male and female children based on their sensation-seeking scores, including boys with low sensation-seeking scores and girls with high scores. This selection strategy may have mitigated the effects of gender since boys tend to score higher on sensation seeking may also take more risks. Our findings also suggest that the gender-based differences previously reported concerning pedestrian behaviour may be driven at least in part by sensation-seeking personality differences in the samples.34

The effect of traffic environment on children’s crossing behaviour

As predicted and consistent with previous studies,23 24 we found statistically significant effects of both inter-vehicle distance and vehicle speed on children’s pedestrian behaviours. We also detected interaction effects: children missed more opportunities for safe crossing, and left shorter delays starting into a traffic gap when vehicles were travelling quickly and at close distances apart than vehicles were travelling slowly and at greater distances apart. The inter-vehicle distance did not seem to influence children’s dangerous crossing when vehicles travelled more slowly at 5 m/s, but children experienced many more dangerous crossings when vehicles were both travelling faster and were spread further apart.

It is unclear why we found children experienced more risk with less dense traffic. However, our findings from the missed opportunity variable—a similar interaction effect, but with exaggerated missed opportunities when traffic was moving quickly and close together—suggest perhaps children, still refining their pedestrian skills through advanced cognitive development, may have been overly cautious in the most challenging traffic condition, when they faced fast and dense traffic. When they faced fast traffic that was more spread, they may have judged the traffic condition as safer than dense traffic, and therefore may ignore the speed of the vehicles and take more risks, sometimes experiencing dangerous crossings.

No matter what the explanation, our findings parallel recent results from Morrongiello et al’s recent study that distance information may play more prominently in children’s pedestrian crossing decisions than speed.20 Confirming evidence is also available in research by Simpson et al in an immersive virtual environment21 and by Connelly et al 19 and Oxley et al 22 in non-immersive situations. The findings also correspond to the developmental psychology notion that children’s spatial concepts develop earlier than temporal concepts.35 36

Limitations and future research directions

Although the present research had several strengths, including a large sample size, use of immersive VR to assess behaviour that would be dangerous to assess in the real world, and recruitment of sub-samples who scored high and low on sensation seeking, we acknowledge limitations also. First, the children we studied were all in middle childhood (10–13 years old), and results may vary with younger or older children, especially given age and developmental effects on children’s pedestrian behaviours.37 38 Second, we focused on children’s crossing behaviour while walking alone. Social influences from peers and adults while crossing streets are both indicated in previous research.39 Finally, the traffic situation in our VR environment was relatively simple, consisting only of traffic travelling on two lanes and at constant speeds and distances apart. Traffic moved fairly slowly, at 5 and 10 m/s. Real-world traffic on Chinese urban roads is far more complex and sometimes moves more quickly. Future research might incorporate greater ecological validity, either in virtual or real-world settings.

Future research might also work to generalise our results cross-culturally. Most previous research on children’s pedestrian safety has been conducted in high-income countries with relatively predictable traffic patterns.40 The Chinese traffic environment is denser, slower and somewhat more chaotic than in most high-income countries,41 and therefore, our results may or may not be generalisable to other traffic environments. Also intriguing is the influence of culture on sensation seeking. Available evidence suggests that the expression of sensation seeking may vary across cultures,42 and therefore sensation-seeking traits might influence children’s pedestrian behaviour in different ways cross-culturally.

Conclusions

Our study suggests that both sensation seeking and traffic environment factors impact children’s behaviour in traffic, and interactions between traffic speeds and inter-vehicle distance impact children’s crossing behaviour. Child sex did not play a significant role in our results.

What is already known on the subject

  • Pedestrian injuries are among the most common causes of death and serious injury to children.

  • A range of risk factors, including individual differences and traffic environment factors, has been investigated as predictors of children’s pedestrian behaviours.

What this study adds

  • Examined the independent and joint influences of individual differences (sex and sensation seeking) and traffic environment factors (vehicle speeds and inter-vehicle distances) on children’s pedestrian safety.

  • Sensation seeking plays a role in children’s pedestrian behaviours, and the impact of sensation seeking on children’s behaviour varied somewhat across traffic flow environments.

  • There were statistically significant effects of both inter-vehicle distance and vehicle speed on children’s pedestrian behaviours; interaction effects were also detected.

  • Vehicle distance information may play more prominently in children’s decisions of crossing than vehicle speed.

  • There were no sex-based differences in children’s pedestrian behaviour.

Acknowledgments

We thank the pupils from the Nantong Wushan Primary School and their parents for their participation. We also thank Principal Liu Zhihe and other teachers from the school for their assistance with this study. Finally, we acknowledge Wang Anni, Hu Huimin and Xue Mingzhu for serving as research assistants for the study.

References

View Abstract

Footnotes

  • HW and ZG contributed equally.

  • Contributors HW and DCS conceptualised the study. ZG and TS developed the VR traffic environment system. FL and JX conducted the experiment, and collated the data. HW conducted the analyses and drafted the manuscript. DCS provided assistance and advice on data analyses. ZG and DCS reviewed and revised the manuscript.

  • Funding This work was supported by the MOE (Ministry of Education in China) Project of Humanities and Social Sciences (grant number 16YJC880072), and was also partially supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development of the US National Institutes of Health under award number R01HD088415.

  • Disclaimer The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Ethics approval Approval for the research was obtained from the Nantong University Academic Ethics Committee and the Primary Education Office of the participating school.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data availability statement Data are available on reasonable request.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.