The aim of this study was to investigate the association between objectively measured neighborhood walkability and children’s PA and the possible moderating effects of neighborhood SES in this association. As the moderating effect of neighborhood SES affects the direct relation between neighborhood walkability and children’s physical activity, the moderating effects will be discussed first.
Two significant moderating effects of neighborhood SES were found. The positive relation between walkability and walking for transportation during leisure was only present in low SES neighborhoods. It is assumed that children living in high SES neighborhoods have more access to motorized transport, and are therefore less dependent on the walkability of their neighborhood; whereas children from low SES neighborhoods may have less access to motorized transport and these children walk more for transportation during leisure time in a high walkable neighborhood, when a lot of destinations are nearby. However, this moderating effect was not found for active transportation to school. It is possible that both children from high and low SES neighborhoods go to school actively, as the prevalence of active commuting in Belgian children is high (59.3% actively commutes to school) , possibly because most of children live close to their school . Therefore, children’s active transportation to school is probably independent of the neighborhood walkability and SES in Belgium.
Neighborhood SES also moderated the relation between walkability and sports during leisure time. In low SES neighborhoods, children engaged more in sports in low walkable neighborhoods and less in high walkable neighborhoods; whereas in high SES neighborhoods, walkability was unrelated to sports. It has been shown previously that SES was positively associated with sports club membership . As children from high SES neighborhoods are more frequently a member of a sports club; the characteristics of their neighborhood are probably less important in order to be active. Because children from low SES neighborhoods do not always have the opportunity to be member of a sports club, due to high costs  and lack of parental support, they probably engage more in unorganized forms of PA such as active street play (e.g. playing street soccer,..). In an Australian study, safety, living in a dead-end street and public open spaces were positively associated with children’s play . Also in the USA  and Canada  children were more active in their neighborhood when low street connectivity was perceived (e.g. more dead-end streets). As these environmental factors are mainly characteristics of a low walkable neighborhood, this may explain the negative relation between walkability and sports during leisure time in low SES neighborhoods.
The fact that children from high SES are more frequently member of a sports club , may also explain the positive relation between neighborhood SES and children’s weekend MVPA and between family SES and sports during leisure time. However this relation was not found for weekday MVPA. As children’s weekday MVPA is highly dependent on the MVPA of children during the school day, it could be expected that children’s SES or walkability was unrelated to their weekday MVPA. Furthermore, no effects of walkability or neighborhood SES were found on children’s MVPA. These results show that environmental factors are differently related to specific domains of PA, which argues for investigating the relation between the environment and PA for different domains of PA, rather than investigating this in relation to overall MVPA or total PA .
Different relations between the environment and PA were found across different age groups in the same city ,,. But similar as in Belgian adolescents in the same city, we found a stronger relation between neighborhood walkability and PA in low SES neighborhoods . However, there is no univocal relation between neighborhood walkability and children’s PA. Living in a high walkable neighborhood can be beneficial for walking for transportation during leisure time, but is negatively associated with sports during leisure time in low-SES neighborhoods. Therefore, based on the present findings the positive and promising results from the adult studies cannot be generalized to children -. This raises the question whether the walkability index, should be changed into a “playability” or “movability” index in order to be relevant to explain children’s PA, as it is possible that other environmental variables (e.g. open spaces and dead end streets), are more important. This implies that physical environmental interventions targeted to increase PA in adults may have opposite effects in children, so attention should be paid when intervening in the physical environment. Children must have the opportunity to play outside and to be active in their neighborhood, especially in low SES and high walkable neighborhoods. Therefore, future interventions should focus on economically disadvantaged subgroups. This may be done by providing play space for children from low SES neighborhoods, e.g. by the provision of sport fields or play streets.
Also the relation between environmental parental perceptions and children’s PA needs to be investigated. As parents are seen as the decision makers for their children , it is possible that parental environmental perceptions are more strongly related to children’s PA.
A strength of this study is the use of objectively measured walkability, as objective measures have less measurement error compared to perceptions . Furthermore, the relation between walkability and PA was measured objectively and subjectively. The cross-sectional character of the study is a limitation, as no causal relationships can be examined. Second, statistical sectors were used to define neighborhoods. It is possible that the investigated neighborhood was not the neighborhood that children and parents would define as ‘their own neighborhood’. Also the low response rate of the school principals is a limitation of this study. Furthermore, data were collected in one city, which may limit the generalizability of the findings. Also, children were recruited in schools instead of neighborhoods with varying walkability and SES levels. This led to the inclusion of children living mostly in high SES–low walkable or low SES–high walkable neighborhoods; children living in low SES–low walkable and high SES–high walkable neighborhoods were underrepresented. Besides, it needs to be acknowledged that it is very difficult to point out the exact relation between physical environmental factors and physical activity, because of the strong interaction between environmental, social and individual factors. Therefore, in future research, the moderating effect of other factors (e.g. family type, number of siblings) in the relation between walkability and children’s physical activity should be investigated.