This study examined the associations between individual, social and physical environmental factors and children’s cycling to school and is one of the first studies to take an age-specific criterion distance for cycling to school into account. We found evidence that in children living within a feasible distance to school individual, social and physical environmental factors were associated with cycling to school. However, it should be noted that neighborhood traffic safety was the only physical environmental factor associated with cycling to school, indicating that the contribution of the physical environment to cycling to school was limited.
In the present study, children from single-parent households were more likely to never cycle to school and to cycle to school irregularly. This finding is in contrast with other studies [12, 23] who found that family structure was not significantly associated with active commuting to school. However, in these latter studies no distinction was made between walking and cycling to school, which may explain the different results. Consequently, no univocal decision may be reached, indicating that more research is required on the contribution of family structure to cycling to school.
Consistent with other studies [12, 32, 33] higher levels of independent mobility were related to more cycling to school. Children who were allowed to cycle alone a certain distance from home were more likely to always cycle to school and less likely to never cycle to school. A possible explanation may be that at this age parents’ direct involvement in active commuting to school with their child is less common . In addition, the significant interaction between independent mobility and children’s age, predicting the odds of always cycling to school, indicated that the difference in children’s cycling behavior between children with low and high levels of independent mobility is more pronounced in older children. Furthermore, children whose biking skills were perceived as good by their parents were less likely to never cycle to school and more likely to always cycle to school. This finding is in line with the study of Trapp et al.  who found that parental confidence in their child’s ability to cycle to school played a mediating role in the association between perceived safety and cycling to school. This study also highlighted the need for educational programs focusing on the development of children’s cycling skills. Our study result underpin this need. Additionally, the fact that reported cycling skills, perceived by the parents, were associated with cycling to school leads us to believe that parents do take the cycling skills of their child into account when deciding to allow their child to cycle to school or not.
An association between parental attitude towards cycling to school and never cycling to school was observed. Similarly, in the study of McMillan  caregivers who valued the social interaction for their child on the trip to school had children that were more likely to active commute to school. Furthermore, children with a high degree of perceived (by their parents) behavior control for cycling to school were less likely to never cycle to school, whereas children who are in the habit of cycling to school were more likely to always cycle to school. These findings are in line with a study of Lemieux and Godin  investigating how well cognitive variables predict active commuting to school. It should be noted that in our study child’s perceived behavior control and habit of cycling to school were reported by parents. However, these results still show that cognitive variables are important predictors of cycling to school, which highlights the importance of cognitions within interventions promoting cycling to school.
Consistent with other studies [8, 36, 37], parental and friend encouragement were also important correlates of cycling to school. Children with friends who encourage them to cycle to school were more likely to always cycle to school and less likely to never cycle to school. In addition, children with parents who encourage them to cycle to school were less likely to never cycle to school. Based on these findings it seems that both sources of social support are important for cycling to school and that especially support from friends is needed to always cycle to school. Furthermore, we found that children were less likely to always cycle to school when parents cycle along with their child. A possible explanation could be that parents who cycle along with their child to school find the trip to school unsafe or have some doubts about the cycling skills of their child. So these parents probably won’t let their child cycle alone to school if they are unable to cycle along which increases the probability that their child becomes an irregular instead of a regular cycler.
Children were more likely to always cycle to school if neighborhood traffic was perceived as safe by their parents. The observation that traffic safety contributed to the prediction of cycling to school is in agreement with other studies [8, 21, 23, 34, 38]. Consequently, efforts like driver education, traffic calming and separate bicycle facilities seem of interest. Remarkably, neighborhood traffic safety was the only environmental factor associated with always cycling to school and no environmental factors were associated with never cycling to school. This finding suggests that the contribution of the physical environment to cycling to school within a criterion distance of 3.0 km from school is limited and highlights the fact that interventions for increasing cycling to school should not focus solely on the physical environment, which supports the view of other studies [35, 38]. Furthermore, the present study found that children were more likely to never cycle and less likely to always cycle to school if their route to school was along roads with walking and cycling facilities. This is rather counter-intuitive, since one would expect the opposite. However, roads in Belgium that are equipped with walking and cycling facilities are usually quite busy. Busy roads have been identified as important barriers for active commuting to school [3, 23, 39].
One of the major strengths of this study is the incorporation of an age-specific criterion distance, which represent a feasible distance for children for cycling to school. Furthermore, an ecological approach was used to identify correlates of cycling to school. Additionally, a distinction was made between correlates of never cycling to school and correlates of always cycling to school. Finally, data were collected in a large sample. Several limitations of this study must be considered. Our data are cross-sectional in nature, indicating that causal relationships cannot be drawn. Further, generalization of this study is limited by the nature of the sample comprising children of the 4th to 6th grade only. However, this narrow age range was chosen in order to obtain a homogenous study population. Furthermore, this study relies only on parents’ reports as children’s perceptions were not measured. Additionally, no objective measures of the built environment were assessed. Notwithstanding this latter limitation, the environmental questions used in the questionnaire have been validated .