The purpose of this study was to examine the reliability and validity of a questionnaire to measure child- and parent-reported psychosocial and environmental correlates of physical activity and screen-based behaviors among Chinese children in Hong Kong. The results suggest that the test-retest reliability was generally acceptable for individual items in both child-report and parent-report questions, as well as the emergent scales which included the self-efficacy scale, the family support scale, the peer support scale, the rules and guidance on screen-based behaviors scales, the perceived neighborhood safety scale, and the social environment scale. These scales showed evidence of convergent validity for most of the items; apart from one item in the child-reported self-efficacy scale and two items in the physical activity support scale, most items loaded onto factors within the scales. The criterion validity analysis also provided psychometric support for using the instrument to measure psychosocial and environmental correlates of physical activity among Chinese children. The results warrant further research into the development of more robust measures, especially peer influences on physical activity and self-efficacy specific to screen-based behaviors.
Since there is no existing assessment scheme specifically developed for Chinese children, most of the items incorporated in this study were modified from published studies, where possible, in the attempt to cover most of the domains identified by ecological theory to be important. Self-efficacy has been consistently reported as a strong correlate of physical activity among children . Self-efficacy measures in the current study were adapted from the existing items specifically designed to examine children's confidence in finding and creating an environment to support their physical activity . Ryan and Dzewaltowski have found that these specific components of self-efficacy had the strongest associations with physical activity, compared with other types of self-efficacy, e.g. overcoming barriers to physical activity, among 11- to 13-year-old American children . However, the results from EFA in the current study showed that one of the five self-efficacy items ("I can walk to school instead of having to wait for a ride") did not seem to be part of the construct and thus stood alone in analysis. Whether or not being an active commuter to school is an important aspect of self-efficacy for participating in physical activity among Chinese children warrants further investigation. However, validity analysis showed that there was no correlation of this specific item with physical activity (data not shown), whereas, a positive association was found between the self-efficacy scale generated by taking the average of the other four items and physical activity (r = 0.25). It is therefore prudent to exclude this item in further studies examining self-efficacy and physical activity in Chinese children. Furthermore, as the self-efficacy scale in the current study is only relevant to children's physical activity, future research should aim to develop similar construct specific to screen-based behaviors.
It has been suggested that the support of family and peers are important correlates of physical activity among children [6, 34, 35]. Measures of family and peer support used in this study were adapted from the measures developed for use with Australian children . The kappa statistics in the present study showed substantial test-retest reliability for most of the individual items, and were comparable to the original measures developed by Hume et al. . Two of the items "Being physically active with grandparents" and "Receive encouragement from a family member" did not load onto the same constructs as the other items or had low loading and therefore was excluded. This was possibly due to relatively low report rates in these two items. The finding that the items on family and peer support reflected two different constructs was similar to what have been found in Iranian girls . In the current study, the peer support scale showed lower Cronbach's alpha (0.50), which was likely due to the limited items in this construct. However, the peer support scale was associated with children's MVPA, whereas the parents' support measure was not. As peer support is closely related to physical activity in school children, especially as they approach adolescence [4, 6], further research is needed to develop and test a broader range of peer support questions relating to Chinese children. The home physical activity environment consisted of newly-developed questions in the current study. Although no association with a similar construct has been found in previous studies with Caucasian children's physical activity, this factor was associated with MVPA in Hong Kong children.
Three measures of neighborhood-level influences on children's physical activity were adapted in this study from existing instruments (i.e., availability of sports facilities, safety concerns, and the social environment). Similar constructs have been extensively examined among Caucasian youth in literature [4, 6, 16, 17, 36–41]. In the current study, the Chinese children were able to consistently report on the available sports facilities in their neighborhood. The finding that the safety scale and the social environment scale showed good internal consistency was consistent with previous studies [21, 28] and emphasizes that a range of questions should be included in order to assess the different aspects of the local environment.
Unlike the correlates of physical activity, studies focusing on correlates of sedentary behavior, in particular, screen-time among children, have been limited . In this study, the test-retest ICC values of the individual items assessing parental rules and guidance in relation to children's screen-time were comparable to the original measures developed by Salmon et al. , and the overall rules scale and guidance scale showed substantial reliability. However, unlike the study in Australian children showing that parents' rules prohibiting TV viewing during meal times was negatively related to children watching TV > 2 hr/day , the rules and guidance scales were not associated with screen-time among Chinese children. It has been shown in a recent review  that two of three studies examining correlates of screen-based behaviors found associations between parent TV time and their child's sedentary behaviors, A similar positive association was also observed in the current study between parental modeling of TV and children's screen-time. It was also interesting to note that family support for physical activity was negatively associated with children's screen time in the current study. These results support the notion that sedentary behavior and physical activity may have their own determinants . Identifying more measures is necessary in future research to examine the correlates of sedentary behaviors among Chinese children.
Strengths and Limitations
This study has several implications for practice. The questionnaire developed in this study provides a useful tool for health professionals, researchers, and school policy makers to examine the multiple correlates of physical activity and screen-time among Chinese children. Based on a social ecological framework, the questionnaire could provide an overall understanding of psychosocial and environmental factors that may shape and change children's physical activity behavior. Further, this questionnaire can provide information on the relative contributions that these factors make to children's overall physical activity levels; thus informing the most to least important factors to target in physical activity promotion initiatives. This questionnaire can also be used in other Asian countries which share similar cultural and environmental characteristics and manifest similar health problems among their inactive youth population.
The results of this study also have implications for tailoring the interventions to increasing lifestyle physical activity among Chinese children. Although correlation analyses were not the focus of this study, validity results showed some evidence of associations with physical activity. For instance, self-efficacy continues to be a significant correlate of children's physical activity as consistently shown in previous studies . Children who reported more support from their peers seemed to be more physically active; although as this was a cross-sectional association, the reverse association could also be true. Family support for physical activity might also be important for children to reduce their screen time. These findings indicate that multiple factors, including individual, social, and environmental aspects, should be considered in order to develop the appropriate behavioral change strategies among Chinese children.
There are several limitations with this study. Firstly, a small number of emergent scales, e.g. peer support, had relatively low Cronbach's alphas. There were only two items relevant to peer influences on physical activity, which inevitably limits the ability to adequately measure this construct. Future research should aim to develop more robust measures of this construct. The other research challenge, which is similar to previous studies, relates to the definition of a neighborhood . This was apparent among the children when asked about access to sports facilities in their community. In the present study, children were guided to think about a "neighborhood" as being the area within a 20-minute walk or drive from their home in order to help them recall the geographic area that they are familiar with. However, it may be that the recreational destinations that children usually go to for physical activity may not necessarily be within the defined area . Future research exploring the neighborhood environmental correlates of physical activity among Chinese children should consider using objective measures in addition to self-report measures.