Background
The rising prevalence of overweight and obesity among children is a public health issue of global importance [
1,
2]. Population studies conducted in Australia, Canada, Ireland, Italy, Malta, Mexico, Slovenia and the United States have classified over a quarter of those measured across the 5-18 year age range as overweight or obese [
3]. Alongside these epidemiological studies has been a large body of research examining the determinants of unhealthy weight gain. High intake of energy dense foods [
4], low levels of physical activity [
5] and prolonged screen viewing [
6] have been identified as causal factors with the potential to be modified. This body of research has contributed to an understanding of the lifestyle patterns that are developed during youth, which may be carried into adulthood, and the risk of early chronic disease development [
7]
The development of effective policies and programs to increase the time children spend in physical activity on a regular basis, and to reduce television viewing and other modifiable sedentary pursuits, requires an analysis of the psychological, social, environmental and economic factors that influence these behaviours. Studies among children and youth have investigated the causes and correlates of total physical activity time [
8,
9], participation in sports and organised activities [
10] and time spent in sedentary pursuits [
11]. The family environment has been consistently identified as a significant influence on both physical activity and dietary intake [
12,
13].
It has been reported that the level of support offered by parents and whether they are role models by their own levels of participation, can influence the time children spend in physical activity or screen viewing [
14]. Parents may determine children's exposure to a number of factors that are enablers or barriers to physical activity, such as money to finance the costs of participation, sports and exercise equipment, transportation to attend activities, and availability of television and other screen activities [
9,
14,
15]. Consequently, given parents' central role, a number of interventions addressing physical activity, sedentariness and overweight and obesity in children have sought to engage parents intensively [
13,
16].
While there is clear evidence about the significance of parental influence, there has been limited analysis of the factors which affect the nature and extent of this influence on child behaviours, including psychosocial, environmental, geographic and economic factors. Further, few studies have examined how parents' influences vary according to children's age. The aim of this study was to investigate how parental self-efficacy and perceived barriers are associated with children's physical activity and screen viewing time. Secondly, the study examined how these relationships differ according to children's age, and household socio-economic and demographic characteristics.
Methods
Design and setting
A cross-sectional survey of children attending preschool and long day care centres, primary (Kindergarten, Grades 2, 4, 6) and high school (Grades 8 and 10) was conducted in the Hunter New England region of New South Wales (NSW) Australia. The study was approved by the Hunter New England Area Health Ethics Committee, the NSW Department of Education and Training and the NSW Catholic Education Commission and data were collected between February and September 2007.
Participants
Parents of children from 16 preschools and 24 long day care centres, randomly selected from a list of licensed childcare centers were invited to participate. Child care centres with < 20 children and those catering for children with special needs (e.g., autism, blindness) were excluded. Thirty-five primary and 35 high schools in total were also randomly selected from each education sector (Government, Catholic and Independent) proportional to the number of students enrolled in that sector. Schools which catered for children with special needs and in remote locations were excluded. Within each school, two classes were randomly chosen from each of the Grades being surveyed and all students invited to participate. Written consent by children and their carers was a requirement for participation.
Questionnaire items
The survey completed by all parents measured perceived barriers and self-efficacy to influence their child's physical activity participation. The physical activity and screen viewing time of preschool and younger children was reported by parents, whereas older children self-completed these measures in a separate survey. Demographic information collected from parents included the child's sex, date of birth, school year, residential postcode, household income and maternal education. Household income was chosen as an indicator of socioeconomic status (SES) [
17] while residential postcode was used to classify children living in an 'urban' or 'rural' locality.
Barriers and self-efficacy
Parents were asked to what extent they agreed with five statements that described barriers which restricted their child's participation in physical activity. These were derived from the parent survey used to evaluate the
VERB child physical activity campaign in the United States, in which they were found to have acceptable item-response and good test-retest reliability [
18]. The statements referred to issues concerning availability of transportation, opportunities for activities in the vicinity of their home, the cost of activities, time constraints of parents, and availability of activities that the child likes. For each statement parents were asked to rate their agreement on a 5-point Likert scale (strongly agree to strongly disagree).
For self efficacy parents were asked to rate their confidence to influence their child's physical activity in a series of challenging situations, which were identified from focus group consultations with parents undertaken by the NSW Health Department (unpublished document). The following eight situations were included: parent does not have much time; the child is engaged in screen time activities; the parent feels stressed; the child does not have a friend to play with; the parent cannot think of activities to suggest; the child is not interested; the parent is not able to participate in the activity, and; the child's preferred activity is expensive. Parents were asked to indicate whether they were very confident, confident, a little confident or not confident in each of these circumstances. Factor analysis showed only one underlying factor for this set of questions in the study sample with good internal reliability (Cronbach's Alpha = 0.88).
Physical activity and screen viewing time
The physical activity questions measured the usual time each spent in organised and non-organised activities on a daily basis while the screen viewing questions measured the usual time spent in a range of screen activities (i.e., TV, recreational computer use, electronic games) each day. The parent report measures used for preschool and younger children were drawn from the NSW Population Health Survey [
19] and were adopted based on research supporting the accuracy of parent proxy reports of physical activity among children under 12 years [
20]. The self-report physical activity questions used with older children have been reported to have acceptable retest-reliability and validity (compared with fitness tests) among 13-15 year olds [
21]. Previous evaluation of the screen viewing questions completed by older children has shown these measures to have very good re-test reliability in 11-15 year olds [
22].
Statistical analysis
Separate analyses were undertaken for preschoolers (attending preschool or long day care centres), younger (school grades K, 2, and 4) and older children (school grades 6, 8, and 10) Demographic covariates were: child's sex; maternal education, categorised as lower (not completing high school), intermediate (completing high school or a vocational diploma), or higher (completing a university degree); residential locality (urban or rural), and; household income (< $40,000, ≥$40,000-$100,000 or ≥$100,000).
For questions about barriers to physical activity the "strongly agree" and "agree" responses were combined, as were "strongly disagree" and "disagree", to create binary variables for each item. The neutral option was collapsed with the answer option that indicated that the respondent did not report a barrier for that item. A 'total barriers' score was derived by summing the number of barriers reported.
Similarly, response options for "not confident" and "a little confident" were combined, as were those for "confident" and "very confident" to form binary outcome variables for the analysis of individual self-efficacy questions. A self-efficacy to influence physical activity scale score was also created from parent responses on the eight questions. Higher scores represented greater self-efficacy among parents to influence their child's physical activity. Because scores on the scale were highly skewed, a categorical variable was then derived splitting the sample into three groups; scores ≤ 25th percentile on the scale were assigned to the low self-efficacy group, scores >25th but <75th percentile were in the moderate self-efficacy group and those with scores ≥ 75th percentile were assigned to the high self-efficacy group. The distribution of parents across these three groups was approximately equal.
A categorical variable for physical activity participation was generated because the data were not normally distributed. Both organised and non-organised activity were categorised to reflect no activity (0 minutes/day), low to medium activity (1-59 mins/day) and high activity (≥ 60 mins/day). Children were also classified according to whether the sum of their organised and non-organised daily activity equalled the physical activity guidelines issued for children, which are ≥ 60 mins/day for younger and older children [
23] and ≥ 3 hours per day for preschoolers [
24]. Screen viewing time was summed and categorised according to guidelines (i.e., <2 or ≥2 hrs/day) [
25].
Bivariate analyses were conducted using Chi-square while logistic regression methods were used in multiple variable analyses. All analyses were adjusted for clustering within school using STATA version 10.0 complex survey commands. The significance level was set at 5%, however, multiple comparisons within significant multi-category independent variables were Bonferonni adjusted for the number of comparisons to reduce the likelihood of Type I error.
Discussion
The findings from this population-based study have shown that parent reported barriers and self-efficacy are significantly related to children's physical activity and screen time. Our analysis of the socio-demographic distribution of these factors, and their relationship with the organised and non-organised activity of children, as well as total activity and screen time, provides valuable insights for developing strategies to address physical activity and sedentariness among children of different ages.
The findings here build on past research which has found that parental support and the family environment play an important role in the physical activity participation of children [
9,
26]. The cost of activities and lack of opportunities in the neighbourhood were the two most common barriers to their child's physical activity that parents reported. While qualitative studies have also identified these issues as barriers that parents face [
27,
28], the present study has shown that these are reported more frequently at the population level than lack of time, issues with transportation or children disliking physical activity. Further, this study has revealed that cost and lack of opportunities are reported far more frequently by low-income parents, providing useful evidence for the design of physical activity strategies for low SES families. This builds on recent research in Australia [
29,
30] and the United Kingdom [
31] indicating that household income is related to choices and level of expenditure on children's physical activities. Another insight related to structural barriers to physical activity was that lack of opportunities for physical activity and issues with transportation were more frequently reported as barriers for families in rural than urban locations, highlighting the importance of addressing these issues for rural children.
Levels of self-efficacy have been identified as a correlate of physical activity participation by school based adolescents [
32], but information about parental self-efficacy to influence the physical activity of children or youth is scarce. Consistent with the barriers frequently reported by low income parents, those in the lowest category of household income most often reported lower self-efficacy. Low self-efficacy due to the expense of child physical activity was twice as prevalent among low income households, compared to those with higher incomes. Interestingly, parents of older children more frequently reported low self-efficacy to influence their child's physical activity, compared with parents of younger children, which is consistent with a recent Australian study [
33]. The present findings suggest that practical communication and planning strategies targeted to assist parents of older children could be beneficial.
The multivariable analyses in this study found that both parent reported barriers and self-efficacy were associated with the likelihood of younger and older children undertaking recommended amounts of physical activity, while for preschoolers only parental self-efficacy was related to meeting physical activity guidelines. It was notable that location of residence was independently related to levels of participation among older but not younger children, with those living in rural areas less likely to be inactive. This was an unexpected finding and suggests that physical activity may play a greater part in the recreational activities of older children in the rural environment.
Consensus about the association between family SES and children's physical activity participation is yet to be reached, with some cross-sectional studies reporting lower participation levels among more disadvantaged children [
34], and others not finding this relationship [
35]. The present study found that, while low income parents reported more barriers and lower self efficacy to influence their children's physical activity, parental barriers and self-efficacy were related to participation levels independently of family income. This underlines the importance of addressing psychological, social and economic factors that affect the ability of parents, at all levels of SES, to enable physical activity participation by their children.
There was an opportunity in this study to examine the association between parents' reported barriers and self-efficacy and children's participation in both organised and non-organised activities. Similar to the finding reported by Heitzler et al [
18] in their study of 9-13 year olds in the United States, parent reported barriers were inversely related to participation in organised physical activity but not non-organised activity. In the present study this relationship was evident, albeit in bivariate analysis, in children across a wider age range. On the other hand, parental self-efficacy was related to time spent in organised activity by younger and older children, but no association was found within preschoolers. For older children, organised sports and recreation activities not only present valuable opportunities for physical activity, but also enable the development of team work skills, leadership attributes and self-confidence [
36]. If these benefits are to be available to a wide range of children this study indicates that sporting organisers need to address barriers to their programs that prospective participants may face, particularly cost, local availability of facilities and access to transport.
A growing body of research shows that sedentary behaviours, particularly screen viewing, are related to the risk of unhealthy weight gain independent of physical activity participation among primary school aged [
37] and high school students [
38]. There is, however, relatively little analysis of the correlates of sedentary time in children [
15,
39]. Here we found that parents with high self-efficacy to influence their child's physical activity were less likely to have children (across all age groups) who exceeded the guidelines for screen time. Hence, when parents report difficulty in influencing their child's physical activity the likelihood of their child exceeding screen time recommendations appears to be higher. These findings are consistent with qualitative research which has found that parents of adolescents struggle to control their child's screen time [
40]. As in previous studies [
41], higher levels of maternal education were also associated with limiting children's screen time, among preschool and younger children in particular. Strategies to promote understanding of the guidelines for screen time, the risk of sedentariness, and techniques for facilitating more active recreational activity need to be considered as a focus for parents with low educational attainment.
Strengths of this study were the large, random sample of parents and children who participated, and the broad age range of the children. Detailed information was collected about parents' barriers and self-efficacy, with multivariable adjustment for a range of potential confounders of the associations between these variables and child physical activity and screen time. Among the limitations of the study were its cross-sectional design, which prevents analysis of the pathways of causality between the study variables, and the use of self-report rather than objective measures of physical activity participation. Further, while the survey response rate was equivalent or superior to other population surveys of children and parents [
15,
18] it allowed some scope for non-response bias. For the analysis of the preschool sample the most recent guidelines for physical activity participation by Australian children up to 5 years of age were used (ie, ≥ 3 hours per day), but these recommendations have not been adopted internationally. Given the topics investigated by the surveys, respondents may have been affected by social desirability bias, however there was effort to minimise this through the use of forced choice items and self-administration [
42]. Finally, the scale that was used to measure self-efficacy had good internal reliability, but its test-retest reliability has not been evaluated.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
BS contributed to survey design and data analysis and undertook manuscript preparation. AG conducted data analysis and contributed to manuscript preparation. LH undertook survey design, facilitated data collection and contributed to data analysis and manuscript preparation. LK contributed to survey design, facilitated data collection and was involved in data analysis and manuscript preparation. LW contributed to survey design, facilitated data collection and was involved in data analysis and manuscript preparation. AM contributed to data analysis and manuscript preparation. All authors read and approved the final manuscript.