This study found that active school travel was mainly related to a combination of parent’s perceptions, socio-demographic and environmental variables, even after adjusting for distance to school. More specifically, we observed that higher rates of active school travel were found in schools where parents perceived low levels of children’s resistance to active school travel, were less concerned about safety and inclement weather and, unexpectedly, perceived the route as having more environmental challenges to navigate (suitability). We also found higher rates of active travel at schools with a higher percentage of Hispanic students and where roads and pathways to school were more pedestrian and cyclist friendly. The three more significant correlates for active travel (p < 0.0001) from the model were parents’ perception of children’s resistance, parents’ perception of safety and weather and the percentage of students who were Hispanic.
Parents’ perceptions of children’s resistance seem to play a key role on the decision of walking or cycling to school in the current study. Similar results have been reported in 5–6 year old  and 9–11 year old [22, 23] Australian children. In the first study with younger children, parents’ perception of their child’s lack of “energy” was correlated to active school travel. In the studies with older children, the parents’ perceptions that their child doesn’t like to walk was negatively associated with active school travel while perceptions that the child preferred to walk was positively associated. The child’s own perceptions and decisions have also been reported to be essential in the decision-making process leading to his/her engagement in active school travel . Although items about child resistance were included in both parent and child surveys’ questions about barriers, it only emerged as a separate factor in the parent survey. And because the parent and child survey data were not linked, it is difficult to distinguish whether it is actual child resistance or parent perceptions of child resistance that is having the greatest impact on active school travel.
While our exploratory factor analysis of parents’ perceived barriers combined issues of safety and weather, with the final model showing them to be inversely associated with active school travel, the literature has generally looked at these issues separately. An association with safety has been regularly evidenced in the literature with parents reporting concerns about traffic safety (e.g., dangerous street crossings), poor pedestrian access to school (e.g., missing or incomplete sidewalks), and crime threats (e.g., bullies) [2, 10–12, 24]. A qualitative study focused in the parents’ decision making process of their children’s school travel mode indicated that the primary decisions were related to safety issues (e.g., traffic, strangers) . However other studies have shown those children’s [26, 27] and adolescent’s  perceptions of safety are generally unrelated to whether they walked to school. Hence, safety perceptions from parents and children seem to diverge. These divergent perceptions were illustrated in a study by Olvera et al., which showed that children perceived their neighborhood safer than their mothers . Lorenc et al. reported that children and young people would like to walk and cycle more and be more independently mobile, but were restricted by their own and their parents’ concerns about safety .
Similarly, the literature shows an association between active school travel and weather. Lorenc et al.’s literature review of active school travel correlates identified three different studies assessing the influence of weather . They reported that bad weather was widely regarded as a disincentive to walking or cycling. Similarly, two longitudinal studies have shown seasonal variability, especially for cycling, supporting the influence of weather on patterns of commuting to school among youth from Norway  and US . However, findings have not always been consistent; others in the US and Australia have found that parents’ perceptions of weather and objective weather assessment related very little to active school travel . Understanding that parents and children are hesitant to walk in rainy or very cold weather should be used with knowledge about local seasonal weather patterns to inform practitioners and interventionist about the best time of year to promote active travel.
The final model also indicated that schools with a higher percentage of Hispanic students had a higher percentage of active travelers, but percentage of African American students was not a significant predictor. Findings are difficult to compare to previous studies as the literature has shown mixed results. A 2008 review concluded that minorities (Hispanics and African Americans) were more likely to engage in active school travel . However, studies published since that review have often found no significant association between active school travel and Latino ethnicity [9, 33–37] or African American race [10, 34–37]. Other studies have even shown ethnicity and race to be associated with lower rates of active school travel [9, 37, 38]. The relationship between ethnicity and active travel seen in the current and earlier studies is hypothesized to reflect cultural norms in which Hispanic adults are more likely to walk for transportation and leisure . However, Mendoza et al. observed that level of acculturation was inversely associated with active school travel among Latino children . Hence the lack of association seen in more recent studies may reflect greater acculturation of those samples. In the present study, level of acculturation was not assessed, so it is impossible to understand to what degree that may explain the association observed. McDonald et al.  were able to determine that a major factor in these relationships between ethnicity and race with active school travel were due to distance between home and school, with a greater percentage of Hispanic (35%) and African American (22%) children living within 1 mile of school compared to white (16%) children. This may help explain why in our final model, which controlled for distance, race was no longer a significant predictor of active school travel. The negative association observed in the bivariate analysis may have been a reflection of African American children living a farther distance from school. Another underlying factor identified by McDonald et al. influencing this relationship between ethnicity/race and active school travel was income . Income is commonly studied socio-demographic variable, and generally, studies indicate that children from a low socio-economic status (SES) background are more likely to walk or bike to school than children with a higher SES [3, 34, 37, 40, 41]. In the present study, household income was not assessed, but the percentage of students receiving free or reduced lunch (an indicator of SES at the school level) was not significantly related to active school travel. While this school-level variable is not a perfect assessment of income, Su et al. were able to demonstrate a positive association with active school travel using a very similar school-level indicator . This interplay of ethnicity, race, acculturation, and income is a complex issue that warrants additional exploration in future research.
Overall the findings suggest that future active school travel intervention efforts should incorporate strategies to work with parents in order to identify the safest route to school or create monitored routes to reduce anxiety about safety issues, improve parent–child communications about active transportation which may help to clarify misperceptions around child’s interest, and to encourage and support active school travel during the most appropriate time of the year. Because distance is so strongly related to active school travel it must be considered when interventions are planned. This could be as simple as targeting only families within a certain distance, or distance to school could be incorporated into a more complex intervention model to help determine the intensity of the intervention for a given family. One model that addresses many of these issues and has been used effectively in a number of communities is the Walking School Bus (WSB) . WSB is a walk-to-school program where children walk to school in groups along a set route (and with set stops along with way), with adults (e.g., parents) essentially serving as the bus driver for supervision. It may be a good strategy to: encourage children to walk together, get parents involved in their children’s active school travel, and be a visible sign of active school travel that models this behavior for the whole community .
Study limitations and strengths
The main limitation of the current study was that child and parent data were not linked and individual child participation was not followed over time. Therefore, all data had to be analyzed at the school-level. Child travel data, which were collected monthly but not matched over time, limit our ability to describe the change in active school travel for individual children. Finally, these data were part of a national evaluation study that occurred over 10 years ago. While there have been national policy initiatives like Safe Routes to School implemented in the interim, funding for such programs has been limited and is estimated to have reached only 10% of schools . The stability in rates of active school travel would suggest that determinants of this behavior have not changed drastically. Therefore, the information presented is believed to remain very relevant for today despite the time lapse between the evaluation and the current analysis.
In spite of these limitations, this study had several strengths. The main strength was the large number of surveys (n = 10,809) that were collected from children over eight consecutive months and the amount of survey information gathered from hundreds of children and their parents from 18 different schools across the US. This information provides a unique national perspective for active school travel patterns. We also used a continuous variable as the main outcome (i.e., active travels per week) ranging from 0 to 10 weekly travels. Most studies include a dichotomized variable of active school travel (active vs. non-active) which may limit their ability to detect the effect certain variables on active school travel, especially given the large number of children that use more than one mode of travel. Finally, we adjusted for distance in all the analysis allowing us to identify factors that relate to active school travel independent of distance to school.