This study examines the factors associated with walking to school and taking part in organised activities. Findings show that income is an important factor in determining the pattern of child activity levels at age 5. In general, walking to school is associated with lower income, while taking part in organised sports is associated with higher income. However, this finding is for children aged 5 and does not appear to be the same as that seen in older children, where more deprived children are reported to be less likely to walk [18, 30]. However, this negative association for older children is not supported by all the literature . For younger children (aged 5), it appears, with increasing income, the type of activity is exchanged from a free but time consuming activity (as most 5 year olds need to be accompanied to school if walking), to a time saving but fee paying activity. Interviews with older children (10-11 year olds) also suggest that physical activity for children in higher socio-economic schools is based around sports clubs, with parents providing financial support and transport to clubs, while, children from lower socioeconomic background have more unstructured play in the street and are verbally encouraged to "go out and do something" . This swapping of activity may have implications when developing new interventions to increase physical activity in children. For example, interventions which encourage uptake of after school sports may have an effect of decreasing walking and overall activity levels may remain stable. Total activity for the individual would need to be considered in any evaluation of interventions aimed at improving any particular aspect of activity.
There are some families who appear to have high levels of general activity as measured by walking to school, playing organised sports and playing with the child. Activity levels were higher in families with evidence of good health behaviours (non-smoking, low BMI of mother, few hours of TV viewing) and those who were more affluent such as having a favourable socio-economic background or living in a 'good' area.
Limitations of the study
However, how factors are associated with activity can not be fully deciphered from this study. For example, walking to school may be inversely associated with distance to school [10, 32] rather than income. Distance to school was not measured in the cohort and is likely to be an important determinant of walking to school. In fact, increasing income may be associated with living in areas that are further away from the school of choice and, therefore, distance to school is likely to be an unmeasured confounder in this study. If distance to school was measured then the volume of active transport (distance and frequency) can also be calculated for each child. Importantly, this study also does not measure activities without the parent. There is no measure of unstructured play in the street (without the parents), and this may be an important source of activity in lower socio-economic children who are not attending clubs and sports. However, in children aged 5 this may have less importance than in older children. This study reports on the parent's opinion of their area which is subjective. There are no objective measurements of the built environment. In addition, this study asks parents how they take their child to school and if they take their child to organised activities. Parents self reported activity may be overestimated and there is no objective validation of the parents self reported activity with the child.
Findings in context with other literature
This study does add to the debate regarding some previously published recommendations, suggesting priority should be given to lower socioeconomic status populations to facilitate environmental change and safety improvements to improve walking to school . We did not find perceived safety of the area influenced walking (similar to Babey et al ). However, this could be because the parents were walking with the children. It is possible that at young ages safety was less of a consideration but it may be an important consideration for parents of older children . Walking to school should be encouraged and targeted in "good" safe areas where working higher income parents live. For example, the Walking School Bus [36, 37] may be most useful for the higher income working white families as organising a parent to walk with the child to school may be a barrier for families with both parents working. However, the design of the neighbourhood, distance to school, availability of pavements/sidewalks and ease of connectivity are all important considerations .
In lower socio-economic areas, policies should also try to improve access to organised sports and clubs to encourage and facilitate involvement from ethnic minority and lower income families. It has been reported that participation in organised clubs/associations is an important part of children's activity, for example, in studies on teenagers this contributes 50-70% of their total physical activity [39, 40]. However, participation would be improved if 'the cost was lower' and as cost has been reported to be a significant barrier to families with daughters and on lower income . Thus, helping financially and improving physical access to sports and clubs could encourage parental support. Parental support is an important influence on the child's activity [41, 42]. Education and playing with the child were found to be linked to both walking to school among higher income families and to involvement in organised sport among lower educational families. Thus, interventions targeted at improving general family health behaviours and support for physical activity appears to be beneficial across all communities. It has been shown  that having parents who value vigorous intensity sports provides the most benefits of activity and reduction of sedentary behaviour, while parents valuing household chores has unexpected negative effects (more TV viewing and decreased team participation). Attitudes vary by ethnicity, education, and number of children , all factors found to be associated with participation in organised sports/clubs within the MCS. Therefore, more research is needed to examine the family health beliefs and attitudes as predictors of the range of physical activities undertaken by children. Previous research  suggests that family based interventions need to accommodate the complex demands of two parent and single parent families and be affordable and varied enough to appeal to a wide range of interests and various ages and development stages of diverse ages of children in the same family.
In summary, children from families that report adoptions of good health behaviours and engage in play with their children, show higher levels of physical activity. Thus, parenting practices appears to be an important part in the activity levels of children. This may include highlighting the importance of making time for playground visits and playing with your child. For example, there appears to be a time trade-off in some higher income families where parents either take the child to organised sports or play with their child but do not do both. Financially subsidising organised sports and exercise might help lower income families participate in more fee paying organised activities and this should be tested in a randomised controlled trial. However, this is likely to need to be combined with working with families to discuss attitudes and support for increased physical activity. Interventions aimed at one pattern of activity should be tested to ensure other forms of existing activity are not reduced.