The present study examined parent-reported PA in different activity levels and the associations with individual and environmental characteristics in children. The mean times that children passed in these activity levels were 57 min/day (6.7 h/week) for moderate PA, and 45 min/day (5.3 h/week) for vigorous PA. The HELENA study reported 71 min/day for moderate and 32 min/day for vigorous PA for self-reported PA across European adolescents . These results are relatively high compared with the current recommendations of 60 min moderate to vigorous PA per day and those reported by objective measures . However, these findings are not surprising because overestimation by questionnaires is a common problem due to recall bias, over-reporting, or by social desirability , particularly in young children [39, 40]. Accelerometers would have been more suitable to record activity continuously and in more detail. They are supposed to be more accurate instruments to measure the spontaneity and variety of the child’s activity . However, this method was too expensive to implement in the present study and therefore questionnaires were used, as usual for large community-based studies [1, 5, 6, 8].
It has been hypothesized that PA in children is determined by a complex interplay of multiple aspects . Using a multidimensional approach, the present findings suggest that male sex, sports club membership, rural region of residence, increased time outdoors and a normal emotional status are associated with a higher level of PA in children aged 10.
Sallis et al.  and van der Horst et al.  reviewed studies on correlates of PA in children and adolescence and showed a positive association between sex (male) and children’s PA in almost all of the 36 studies they examined, which is consistent with the finding of the present study. The reasons for sex-specific differences were discussed from different perspectives: while lots of investigators have focused on socialization in sports or exercise involving family, school or peer groups [43, 44], a study by Vilhjalmsson et al.  found lower enrollment in organized sport clubs among girls and suggested that organized sport may be a key explanatory factor. In the present study, the opportunity to engage in vigorous PA was reduced for children who were not enrolled in a sports club. We also observed a reduced enrollment in sports clubs for girls. Both of our findings provide some support for the hypothesis of Vilhjalmsson et al. However, the study design of the present study does not allow confirming this hypothesis and thus it is important to consider further potential explanatory factors like child’s physical environment. With respect to this factor time spent outdoors was positively correlated with a child’s overall PA. This finding is consistent with the results of Sallis et al. Moreover, the present study demonstrates that the effect of time spent outdoors is most significant for moderate activity levels in summer. The seasonal differences can be ascribed to the nicer weather in Germany during the summer season and agrees with the findings of Conrad et al. . The present findings show that children spent much more time outdoors in summer than in winter and that there is a significant difference in the time spent outdoors between girls and boys, but only in winter.
A neighborhood environment that offers possibilities to be physically active on playgrounds was found to promote PA time among children [42, 45, 46]. Time spent outdoors decreases with increasing community size and urbanization and is higher in neighborhoods consisting of detached houses . In the present study, four study regions were examined which allow comparisons between urban (Munich and Leipzig) and rural areas (Bad Honnef and Wesel), as well as between East (Leipzig) and West (Munich) Germany. Children from Wesel had a greater chance to be engaged in moderate or vigorous activity than children from Munich. Accordingly, children from more rural areas seem to be more active than those living in urban areas, a finding also often observed by other studies [45, 47]. A reason for the increased moderate PA of children from rural regions could be a higher time spent outdoors, as observed among children from Wesel. A comparison between Bad Honnef and Munich did not yield conclusive findings. This is probably caused by the low number of subjects (n = 161) living in Bad Honnef. Differences between East and West Germany, i.e. between Leipzig and Munich, became apparent in winter, especially for high activity levels suggesting that children living in urban areas of East Germany seem to be more active than those in West Germany. Since the climatic conditions between the four study regions are comparable  differences cannot be related to the regional climate. Up to now, studies on psychological factors have demonstrated that PA is positively associated with self-esteem and self-concept . With respect to emotional symptoms an inverse association between PA and depression, anxiety, and shyness was found in children [22, 50, 51]. The present findings support these associations, showing that it plays a more prominent role for high intensity levels of PA. Children with abnormal emotional symptoms probably avoid situations that need “overcoming”, for example, when performing high intensity sports. Anxious or under-confident children are probably less prone to take part in sport class, which we found to be correlated with vigorous PA. Therefore, it would be commendable to enhance childrens’ perceived competence in relation to PA by helping them to overcome the barriers and to go through a series of successful experiences regarding PA .
Some variables, for which a link to children’s PA was hypothesized, showed no association. For example, none of the socioeconomic variables contributed to the regression models. Parental education and parental income did not show any effect, except for one subgroup of vigorous PA that was moderately influenced by income. Many other studies have also reported inconsistent associations for parental education and income with PA [9, 10, 53]. However, comparison of findings with other studies is hampered by the fact that different measures for socio-economic status have been applied, including education, income or occupation . There was also no correlation between parental BMI and PA in the present study. As previously reported , no decisive links were observed between the children's BMI and status of puberty and PA. Obesity and overweight were low in the study population, 2.4% and 6.6%, respectively, which may limit conclusions on this subject. Other studies revealed inconsistencies between BMI and PA but the majority could not show an association between the two variables . The same applies to studies which examined the association between puberty and PA. Some studies found PA levels varying with pubertal status whereas other studies could not find any association between puberty and PA . Furthermore, behavioral aspects such as TV/computer consumption or the method by which children reach school did not influence the multivariate model. Indeed, in our study, only few children showed an extreme TV/PC consumption of more than two hours per day and, therefore, it is difficult to draw conclusions on this subject. However, it is remarkable that boys watched more television than girls, but at the same time exhibited higher activity levels. Conflicting results emerge from studies on TV/computer consumption in children suggesting on one hand that increasing amount of time in sedentary activities seems to negatively influence PA  while on the other hand, sedentary time is not necessarily an indicator for reduced amounts of PA [15, 57, 58]. The study by Pate et al.  showed that children spending more than three hours of TV/computer consumption per day are more likely to be low active suggesting that there may be a threshold beyond which the impact of TV watching is perceptible on PA behavior .
The results of the present study suggest that physically active children show particular characteristics that significantly differ from low active children. Active children are usually a member of a sports club, spend a lot of time outdoors and do not appear to have emotional symptoms such as fear, unhappiness or frequently headaches. Moreover, they tend to be male and live in rural areas. Though these two factors are not modifiable, they can be used to identify children at risk. Preventive interventions should focus on motivating children to increase their PA with particular consideration for girls, children in poor emotional mood and urban children.
Overall our findings are confirmatory but, in contrast to other studies, in the present survey multiple factors associated with PA in children were considered simultaneously which results in fewer significant associations than did other strategies such as bivariate analysis . Hence, the results of the present study exclusively reflect essential correlates of PA in children. The large sample size is another important strength of this study, as it allows the calculation of precise estimates for several different parameters. Moreover, for the multivariate analysis, categorical data were used which rendered our results less susceptible to outliers. Limitations of this study include the reliance on parent-reported data and the cross-sectional study design. Moreover, there may be additional correlates of PA which were not measured and therefore could not be considered in our study, like family support, parental PA or school-based strategies that encompass physical education, classroom activities, and after-school sports [9, 10, 46]. Interpretations according to the socioeconomic status should be treated with caution as the socioeconomic status of the ten year follow up is above average. There is evidence that PA questionnaires lead to higher levels of self-reported physical activity than measuring it with the help of accelerometers . However, because of the large sample size of our study questionnaires were more suitable. Nevertheless, there is need to determine a valid and inexpensive instrument that can be used in large community-based studies to collect physical activity behavior of children. Furthermore, it is essential to initiate longitudinal studies to identify the direction of causality of potential correlates of PA.