1. See http://www.oecd-ilibrary.org/sites/health_glance-2011-en/02/04/index.html?contentType=&itemId=/content/chapter/health_glance-2011-19-en&containerItemId=/content/serial/19991312&accessItemIds=/content/book/health_glance-2011-en&mimeType=text/html
2. No information on this is currently available for Scotland.
4. For further details see http://www.scotland.gov.uk/About/scotPerforms/partnerstories/NHSScotlandperformance/childhealthyweight
5. For the obesity evidence, see http://www.scotland.gov.uk/Topics/Health/good-places-better-health/Recommendations/Obesity
6. See http://www.scotland.gov.uk/Resource/Doc/302783/0094795.pdf and http://www.scotland.gov.uk/Resource/Doc/346007/0115166.pdf
10. Small numbers precluded further division of the minority group category.
11. Similar results were obtained in alternative models not shown here, which substituted screen time and fast food consumption for (respectively) bedroom TV and eating in a room without a dining area. However, child-friendliness of the neighbourhood proved a stronger predictor of children's obesity in multivariate models than either neighbourhood safety or antisocial behaviour problems.
12. Note that if we included sweep 4 BMI, we would have been measuring change in BMI status in the models. However, we did not have two complete sets of potential risk factors, measured before and after sweep 4. In the interests of simplicity, it was decided to focus solely on the sweep 6 measures of overweight and obesity.
13. These differences are unlikely to reflect methodology, as questions on physical activity were similar in GUS and SHeS. The discrepancy might be due to differences in sample populations and/or measurement error in parental estimates.
14. Although for simplicity we present findings for a measure that includes school-based activity, analysis with a measure excluding school activities gave similar results.
15. This cut-off was chosen to reflect increased obesity risk, while being sufficiently discriminatory (imposing a lower cut-off of say 2 hours would include around half of the sample in the "high screen time" group). The weekday rather than Saturday measure has been selected because of its likely greater implications for reduced time spent on physical activity, and because of its clearer association with overweight and obesity.
16. This measure is best treated as a measure of what mothers thought was desirable, rather than measuring awareness of UK recommendations. Fieldwork was conducted during 2010-11. UK guidelines on physical activity were not published until 2011, although they were based on similar WHO recommendations published in 2010, see http://whqlibdoc.who.int/publications/2010/9789241599979_eng.pdf.
17. Access to a garden was reported by 90% of families at sweep 1.
18. Note that with the increased prevalence of overweight and obesity in children, mothers' perceptions of overweight as a "normal" condition have some justification. With hindsight, it might have been preferable to have used the term "healthy weight" as a response option, rather than "normal weight".
19. Findings for obesity should be treated with caution due to small base samples.
20. Further details provided on request.