healthscotland.com

Sedentary behaviour among Scottish youth: prevalence and determinants

2. Measurement of sedentary behaviour in youth

2.1 Method

The principal data collection instrument was a self-report diary of “free-time” behaviours that school students completed outside of school hours. Because we were primarily interested in behaviours where young people had some element of choice, behaviours in school were not assessed. The diary was based on principles of Ecological Momentary Assessment (EMA) and was divided into two parts. The first involved questions about child-level variables (9 items; “About You”), family-level variables (11 items; “About your Family”), and environmental-level variables (15 items; “About your Home”) that have been hypothesized to correlate with sedentary behaviour and physical activity. The second part was for recording the behaviours and locations that young people engage in each day in their free time. Reliable estimates of behaviour have been found (Marshall, Biddle, Murdey, Gorely, & Cameron, 2003) .

Participants were asked to complete the diary for four days (three weekdays and one weekend day). At 15-minute intervals, participants self-reported (free-response) their main behaviour in response to a single item: “What are you doing now?”. At each interval, participants also responded to the closed-response items “Where are you?” (LOCATION) by selecting one location from a list of 12, and “Who’s with you?”. This last question will not be analysed and reported here due to limitations of space. An example diary page can be seen in Appendix 1.

Data collection days were randomly assigned by weekday and weekend day. For each weekday, 44 time-samples were obtained (one every 15 minutes from 07.00h to 08.45h and from 15.00h to 23.45h). For the weekend day, 68 time-samples were obtained (one every 15 minutes from 07.00h to 23.45h). To control for different levels in school-based physical activity affecting out-of-school behaviour, two items assessed participation in physical education (“Did you take part in a PE lesson today?”) and engagement in moderate-to-vigorous physical activity (“At school today, did you run round or breathe hard enough to make you sweat?”). At the end of each diary day, participants also responded to an additional 10 closed-response items that solicited information about snacking behaviour during the day and events that may have affected diary entries (e.g., weather, injury/illness, etc.).

The behaviours were first coded into 23 mutually exclusive categories. To estimate the time spent in each behaviour category, at each location, and in each social context, the interval-level data were aggregated for each individual (separately by weekday and weekend day) by multiplying the daily frequency of the event by 15 (1 interval = 15 minutes). The weekday data were then aggregated further to produce a mean, in minutes per day (min.day ^-1), across weekdays. The outcome variables for all analyses are min.day ^-1 engaged in 23 categories of behaviour in 12 locations (see Table 2.1).

Sampling took place across 14 local education authorities (LEAs), randomly drawn from the total of 32: Aberdeenshire, Dumfries and Galloway, Dundee, East Ayrshire, South Ayrshire, Edinburgh, Falkirk, Fife, Glasgow, Highland, North Lanarkshire, South Lanarkshire, Perth and Kinross, and Renfrewshire. Schools were then randomly selected from each LEA to be approached to take part in the study. Ratios of secondary to independent schools were maintained in the selection process. From www.schoolsnet.com a list of schools in each of the selected LEAs was printed before random numbers were generated and the corresponding schools selected. If a selected school had less than 300 pupils, the school was discarded for reasons of representativeness, and another randomly selected.

Schools identified were contacted by letter and invited to participate. The aims and expectations of the study were outlined and schools that agreed to take part were sent a pack including three sets of 30 diaries (approximately 75 for distribution and the rest as spares in case of loss, damage etc) and an evaluation form. Useable data were returned by 27 schools.

In Phase 1 (October-November 2002), one class from each of the Year Groups S2 (mean age=13.24 years), S3 (mean age=14.24y), and S4 (mean age=15.18y) was chosen at random by a co-ordinator at the school and each student within the chosen class was given a diary to complete in their free-time. Schools were offered the incentive of £1 for the return of each completed diary. In order to assess, and control for, possible seasonal variations in behaviour a second phase of testing was carried out six months after Phase 1 in February-May 2003. Each school that returned Phase 1 data was asked to select a second set of Year S2, S3, and S4 classes and repeat the study. It was stressed that Phase 2 classes should not include any student who had previously returned data in Phase 1. Mean ages at Phase 2 were S2=13.62 years, S3=14.55y, and S4=15.55y.

Table 2.1 Behaviours and locations assessed using EMA diaries

Of the 1056 participants who returned diaries, 60% were girls. Assuming there is an equal proportion of girls and boys in the school population, our sample reflects a gender bias beyond that expected by chance (Chi-square (X 2) = 38.64, df = 1, p < .01). Of the returned diaries, girls were also more likely to provide complete data (χ 2 (4, N = 1056) = 40.66, p = .0001). Forty participants (4%) returned diaries with completed demographic data but no diary data (i.e., partial responders) and were therefore excluded from further analyses. Partial responders were more likely to be male (χ 2 (1, N = 1056) = 22.14, p = .0001), slightly younger than full responders (13.7 vs. 14.1 y; t (1059) = -2.58, p = .01), and live in a less affluent area ( p = .001), but they did not differ by ethnicity ( p = 0.68) nor whether they found the diary harder to understand ( p = 0.57), compared to full responders. See Table 2.2 for summary data on participants.

Table 2.2 Demographic characteristics of the final sample