The objective of this study was to examine the unique contributions of demographic, individual, social, and environmental factors to MPVA levels among boys and girls aged 10-17-years-old. The results of this study add to growing evidence that suggest that important difference may exist between boys and girls in terms of the correlates related to time spent participating in MVPA. As expected, boys exhibited significantly higher levels of MVPA than girls, although both groups accumulated far less MVPA than the recommended 60 minutes per day (i.e., 21 minutes on average). Using accelerometer data from the Study of Early Child Care and Youth Development, Nader et al.  found that at age 15, adolescents were engaging in MVPA for 49 minutes per weekday and 35 minutes per weekend day. One explanation for the low prevalence in the present study may be the higher MET (metabolic equivalent) value (i.e., four) that was employed to define MVPA. As evident in the literature, there is no single accepted method for assigning MET values or defining MET cutoff points among children [44, 45]. In the present study, the convention of Troiano et al.  was used where moderate intensity equals four or more METs, which takes into consideration the higher resting energy expenditure of children and adolescents .
Among boys, factors at the individual (self-efficacy), social (peer support), and environmental levels (home PA equipment and temperature) emerged as important predictors of MVPA. For girls, factors at the individual and environmental levels were significant; the lower their perceived barriers, the closer they lived to the school they attended, and the more walkable their neighborhoods, the more MVPA girls' participated in. The combination of these variables may reflect the patterning of boys' and girls' respective activities, particularly in the middle and high school age ranges. Boys have been found to take part in more free-time or unstructured PA than girls , activities that often take place outside, with peers, and using equipment from within thee home. The idea of "pick-up" games or neighborhood activities may be more prevalent among boys than girls, and more influenced by one's own confidence (or self-efficacy), the support of peers, the home environment, and the temperature outside. While one might hypothesize that the neighborhood environment (i.e., access to trails and parks) might align with this theory, it may be that these same activities can take place in the home drive-way, backyard, or nearby open field - all venues that would not be captured as unique PA environments via GIS measures.
Previous investigations have shown self-efficacy to be one of the most consistently associated factors with PA among youth [7, 8] and an important mediator between social and environmental variables and PA , although the majority of these investigations have been limited to girls. These findings point out the importance of self-efficacy among boys as well. Likewise, peer support has also been shown to be an important determinant of PA among children and adolescents in several studies [15, 16, 48]. Interestingly, parent support and parents' own PA levels were not found to be important correlates of MVPA among either boys or girls. Similarly, Duncan et al.  found that friend support was a stronger influence of 10-14-year-olds' PA than parent or sibling support. In particular, watching activities was reported as being the most influential factor - a form of support that was considered emotional, rather than instrumental, among these authors. Likewise, Beets and colleagues  found that peers were the only social support provider related to the activity levels of fifth to eighth graders.
Existing research on the availability and accessibility of PA and sports equipment within the home is limited [13, 49]. Typically, the measures included in such investigations are made up of one or two items which assess children's perceptions of equipment availability as opposed to a more thorough inventory of both the quantity and accessibility of items, as was done in the present study. The current findings strengthen the need to continue exploring the influence of the home environment as part of the broader social and physical context in which PA behaviors can take place and to examine differences that might exist by gender. However, it is important to interpret these results within the cross-sectional context in which these analyses took place. It is impossible to determine whether having more PA equipment available and accessible in homes causes an individual to be active or if being active causes homes to be more populated with equipment.
Average monthly temperature had the highest standardized coefficient in relation to boys' MVPA, after adjustment for all other variables. That is, the higher the average monthly temperature, the more observed MVPA among boys; although this relationship was not seen for girls. The region in which this study took place includes sustained periods of very cold temperatures and significant snowfall, variables which assumingly can affect both the quantity and types of activities that youth may participate in. A study by Brodersen et al.  also showed differences by gender between weather-related variables and children's PA. In this previous study, rainfall was negatively associated with activity levels among girls, however, among boys, lower temperatures, but not rain, were positively related to sedentary behavior. Boys preferences for activities such as basketball, baseball, football, soccer, and riding bicycles  and media-based activities such as watching television and playing video games  may be more influenced by variable temperatures as compared to activities that girls may be more involved in. The present findings point to the importance of offering both indoor and outdoor opportunities and facilities so that PA participation can continue to take place throughout the colder months in less temperate climates.
Among girls, barriers related to PA and the distance to ones' school may reflect their ability or enjoyment related to organized activities. In the present study, girls reported higher barriers than boys. The list of barriers that was included reflects obstacles often associated with participation in "traditional" sports such as being chosen last for teams and being embarrassed. Therefore, identifying the salient barriers among girls and helping to identify attainable steps to overcome such barriers may be the first issue to tackle. Living closer to ones' school may facilitate higher participation in school-related activities including intramural sports and after-school activities such as basketball, cheerleading and dance, popular activities among many girls this age [46, 51]. It is also conceivable that parents may be more willing to sign-up and/or transport their children to after-school activities if they live closer to the school that they attend.
In addition, the present study found that the more walkable girls' neighborhoods were, the higher their MVPA, a finding similar to previous research. For example, Norman et al.  found a negative correlation between intersection density (a marker of street connectivity and walkability) and MVPA among girls, and not boys, aged 11-15-years-old. Given the higher MET value that was assigned to moderate PA in the present study, it is unlikely that this relationship reflects walking behavior performed for transport (i.e., to retail stores, school, or friends' homes), which would typically be characterized as light (2.9 METs) or moderate (3.6 METs) . However, it is conceivable that girls in this sample who live in more walkable neighborhoods are more likely to walk for exercise, an activity found to more likely among females than males . More research is needed to further investigate how the design of girls' and boys' neighborhoods might specifically influence their activity levels and the different types of activities they prefer.
Approximately 25% and 15% of the variance in MVPA was accounted for in these analyses for boys and girls, respectively. These results are similar to  or exceed  other studies examining youth PA at multiple levels. The findings support using ecological models that emphasize multiple levels of influence to better understand, as well as to influence PA behaviors. However, it is clear that there may be other important variables that determine youth's participation in MVPA. Taken as a whole, the observations of this study reinforce the need for messages and interventions to consider salient gender differences in the determinants of PA.
There are several limitations to the present study. The use of cross-sectional data does not provide evidence that the variables under investigation are causes of MVPA among youth. Similar to previous research [13, 26], none of the demographic variables were independently associated with PA among boys or girls. The study sample was predominantly white and of higher socioeconomic backgrounds, making it difficult to make comparisons by race, ethnicity or socioeconomic status. As has been documented, minority and low-income youth may be at higher risk than other children for low levels of PA [3, 26]. The contextual factors influencing more diverse populations of youth may be very different from what we described in this research. Additionally, the smaller sample size may mask some meaningful effects that would be statistically significant if a larger sample was available. There are also limitations related to the GIS measures. While these measures can assess the extent to which facilities or resources are available in children's neighborhoods, they do not address the specific features or quality of those resources, including safety, cost, and age-appropriate programs. The density of these facilities within children's neighborhoods was also not included in these analyses. In addition, assessing the environment around an individual's home may not necessarily reflect the facilities that they actually use or other environments in which they are active.
Nonetheless, the results provide valuable information regarding issues to consider when determining programmatic needs or assets and provide direction for future longitudinal analyses. The major strengths of this study were the multiple, diverse measures used for assessing the individual, social, and environmental factors that may influence PA among youth, including objective measurements of the environment and MVPA and the wide age range of youth included. Future studies, including longitudinal analyses and research among more diverse samples, should consider additional variables, such as neighborhood social cohesion, perceptions of and actual crime data, and more detailed measures of the psychological and social factors such as perceived competence and social support related to both free-time and organized activities.