e64 Abstracts / Journal of Science and Medicine in Sport 12 (2010) e1–e232 Australia questionnaire. Of the 1281 participants at baseline, 846 agreed to be recontacted and follow-up interviews were conducted with 368 adults in September 2008 using the same survey instrument. The response rate for the follow-up sur- vey was 73.2%. Relationships between changes in activity levels and perceptions of the environment were examined using logistic regression, results are presented as both unad- justed and adjusted OR for demographics, self-efficacy, social support and residential self-selection. Results: There was no statistical significant associa- tion between increased total minutes of physical activity and perceptions of the environment in unadjusted analy- sis (OR = 1.46, 95% CI 0.94–2.28) or adjusted analyses (OR = 1.41, 95% CI 0.89-2.24). Significant relationships were observed between increases in walking for recre- ation and perceptions of the environment in both unadjusted (OR = 2.27, 95% CI 1.32-3.89) and adjusted analyses (OR = 2.36, 95% CI 1.35–4.11). Conclusion: The observation that favourable changes in neighbourhood environments are associated with favourable changes activity over a 7 year period confirm previous results of cross sectional studies observing that higher activity levels are associated with more favourable neighbourhood envi- ronments. The lack of association between total minutes of physical activity and perceptions of the neighbourhood likely reflects the characteristics of the neighbourhood environment assessed. This work was supported by a Grant-in-Aid (G07B3114) from the National Heart Foundation of Australia. doi:10.1016/j.jsams.2009.10.131 131 Changing association of Australian parents’ physi- cal activity and their children’s sport participation: 1985–2004 J. Dollman University of South Australia Introduction: The socio-ecological milieu of children’s physical activity is changing over time, perhaps causing per- turbations within the causal “web” that explains physical activity behaviours. This may partly explain the equivocal lit- erature in this area. While parental role modelling has gained acceptance as a correlate of children’s physical activity, it is unclear if its relative importance is changing. Accordingly, this study examined associations of parent physical activity and children’s sport participation in 1985 and 2004. Methods: In 1985, children in 10 South Australian schools participated in the Australian Schools Health and Fitness Survey (ASHFS), and children in eight of these schools par- ticipated in a repeat survey in 2004. Sampling methods and administration protocols for the AHFS questionnaire were identical in both surveys. Items relevant to this study were: number of organised sports played in the previous 12 months (“Sport”); and children’s perceptions of whether mother and father exercise at least twice weekly (Parent PA, response options: yes; no; don’t know). Parent PA responses were combined: both active; one active; neither active. ANOVA compared Parent PA on Sport, separately by sex and survey year, among children (9–15y) from the 8 schools common to both surveys (1985:179 girls, 211 boys; 2004: 210 girls, 218 boys). Results: Parent PA differed between surveys (p < 0.0001). In 1985,19% reported neither parent active while 36% reported both active, compared with 32% and 28% in 2004. In 1985, there were no differences (p > 0.05) in Sport between Parent PA categories among boys (both active 2.5 ± 1.7; one active 2.3 ± 1.5; neither active 2.2 ± 1.4) and girls (both 2.1 ± 1.6; one 1.8 ± 1.5; neither 1.7 ± 1.5). Among boys in 2004, Sport was higher in those with both active parents (2.6 ± 1.5) than those with one (2.0 ± 1.5, p = 0.009) and neither (1.6 ± 1.4, p < 0.0001) active, while those with one active parent differed significantly from those with neither active (p = 0.04). Among girls in 2004, Sport was also higher in those with both active parents (2.5 ± 1.7) than those with one (2.0 ± 1.5, p = 0.03) and neither (1.3 ± 1.3, p < 0.0001) active, while those with one differed from those with neither (p = 0.04) active. Conclusions: In the 2004 sample only, children who perceived both parents to be active reported highest sport participation, while those with neither parent active reported lowest sport participation. This underscores the current role of parents as socialising agents for physical activity. More generally, the results suggest that intervention design should be founded on the most recent available evidence of children’s physical activity correlates. doi:10.1016/j.jsams.2009.10.132 132 Childhood obesity - are we there yet? T. Olds ∗ , C. Maher, K. Ferrar, G. Tomkinson University of South Australia Background: The popular media, health experts and researchers in Australia and overseas talk about an “obesity epidemic” with exponentially increasing rates of obesity and overweight. The aim of this study was to examine recent trends in the prevalence of childhood overweight and obesity in Australia and overseas. Method: 41 surveys of childhood weight status conducted in Australia between 1985 and 2008 were reviewed. The stud- ies included data on 264,905 Australians aged between 2 and 18 years, including raw data on 70,758 children. Children were classified as overweight or obese using the criteria of Cole et al. (2007). Prevalence estimates were adjusted for age, and plotted separately for boys and girls against year of