Several reported neighborhood environment correlates were found for child physical activity in the neighborhood, supporting the idea that stronger correlates will be identified for behaviors if the contexts of the correlates and behaviors are aligned [2, 26, 27]. Higher probability of children being reported active in the neighborhood was related to lower reported street connectivity (e.g., more cul-de-sacs), as found in another recent study . Low connectivity reduces traffic volumes, providing safer neighborhood places to play. This finding contrasts with higher rates of transport-based physical activity among adults and children living in neighborhoods with higher street connectivity, such as more child active travel to/from school, particularly in low traffic areas . Because adults are not likely to play in the cul-de-sacs and alleys like their children, these findings seem to present a tradeoff between child activity and adult activity based on physical environment. These findings together highlight the importance of examining specific types of physical activity (and subgroups within the population such as children versus adults) in relation to the environments in which the activity does or could occur, to know where best to target limited resources for public policy or urban planning projects. Creative planning solutions could allow parents and families to engage in active transportation to work, school, and other destinations allowed by better street connectivity while providing opportunities for children’s unstructured play in the neighborhood either in streets (e.g., traffic calming strategies) or by providing other proximal spaces for active play (e.g., more playgrounds).
As with adults’ recreational physical activity , we found that aesthetics were related to more child neighborhood physical activity. Favorable aesthetics may improve the enjoyment of being active in neighborhoods. However, Limstrand, 2008  summarized the limited literature on this topic and found that two out of three studies reported no association between aesthetics and children’s physical activity level. Another review found little relation between vegetation and children’s reported activity . Attractive buildings and gardens may have provided a sense of order for our particular sample of parents, making them more comfortable to let their children play and be active outside.
More frequent children’s park activity was associated with better safety from crime, aesthetics, walking/cycling facilities, and suitable play area availability. Prior evidence highlights the importance to children’s physical activity of having better accessibility to play areas [19, 32]. Regarding aesthetics, we found previously that neighborhood aesthetics were perceived as higher by parents of children with more frequent park-based physical activity . This prior study however found no relation between neighborhood safety from crime and park activity. Differences in sample characteristics and the types of urban areas and neighborhood environments may explain this discrepancy.
Proximity to play areas was the only correlate of total physical activity in the present study, and it applied to parent-reported child overall activity as well as to child accelerometer-based MVPA (both scoring methods). Others have found that play area proximity is consistently related to physical activity (as reviewed in ). Perhaps this association with overall physical activity reflects the fact that, for children, such physical activity is more discretionary than other physical activity in which children engage (e.g., school-based physical activity) or that higher levels of physical activity are only generally reached by children routinely using such play areas. Having few significant environmental correlates of the overall measures of physical activity is a finding consistent with previous literature . Environmental influences are expected to be specific to domain of physical activity, such as transport or recreation [26, 33], or physical specific location, such as neighborhood or park . Thus, total physical activity measures may underestimate the importance of individual neighborhood environmental factors, although are useful given the association with health outcomes. Measuring physical activity in specific contexts will provide more useful information to those agencies and programs attempting to institute environmental changes to increase activity in these contexts.
Notably though, access or proximity to recreation areas was related to all physical activity outcomes in the present study, whether location-specific or total, parent-reported or objectively measured. This impressive consistency in a literature characterized by inconsistency  suggests that proximal play areas could be a powerful influence on child physical activity, and that this may be a useful focal point for cities that are developing more active-friendly neighborhoods. Recreational physical activity is the dominant domain of activity for children, and children typically have very low levels of physical activity while indoors , so it is reasonable that having places to play near the home would emerge as an important correlate of physical activity. Conversely, lack of accessible places to play could be a very strong barrier to children's activity.
Additional analyses are needed to examine whether objective or perceived built environment, or what combination of these, are more strongly related to children’s physical activity and where that physical activity occurs. The length of time spent being active in specific locations would also provide added value. A longitudinal study design would make causality inferences between the predictors and activity outcomes more feasible to determine, as would natural experiments in which children’s behavior is examined before and after environmental changes. The present study did not include objective location or environment data such as what can be measured using GPS and GIS, which would allow us to objectively measure actual distances to play locations relative to amount of activity, rather than relying on parental reports, some of which had low internal consistency.
Among study strengths, children from a diversity of neighborhoods were recruited from two distinct regions, and virtually all measures demonstrated good retest reliability. However, most children were from relatively affluent families, with representative but limited racial/ethnic diversity. The recruitment rate of this study was lower than in survey-only studies, likely due to the added burden of an office or in-home visit (to measure anthropometrics) and accelerometry, leading to self-selection bias of higher SES families. However, this higher SES does not seem to have resulted in abnormally high activity rates relative to a national sample [35, 36]. Parental perception was a limiting factor in the assessment of the neighborhood environment, and measurements such as distances to parks and recreation areas were subjective and potentially prone to error. These analyses were limited to recreational activity and not transport activity, a potentially significant limitation given the substantial role that active transport by children to school can play in some neighborhoods and for their overall physical activity.