NSW School Students Health Behaviours Survey

The 2017 School Students Health Behaviours Survey represents the twelfth survey of a series that commenced in 1984. It captures information on a wide range of health behaviours, including smoking, alcohol consumption, and illicit drug use in New South Wales. This section describes the methods of data collection and analysis.

Sample selection

The target population was all students in Years 7-12 enrolled during the period February to December 2017 in New South Wales. Schools with fewer than 100 students were not included in the survey. Language schools were also excluded from the sampling frame.

The survey used a 2-stage probability sampling procedure: schools were selected first; students within schools were selected second. Schools were stratified by the three sectors (Government, Catholic, and Independent) and randomly selected within each sector. The sampling procedure ensured the distribution of schools among the three sectors was reflected in the sample. Two samples were drawn: junior secondary (Years 7 to 10) and senior secondary (Years 11 and 12).

The target school sample was 126 secondary schools in 2008, 2011, 2014, and 2017. In 2017, to try and achieve this target, 764 schools were approached (437 in 2014) and 94 schools agreed to participate (112 in 2014), giving an overall response rate of 12.3 per cent (26% in 2014). The survey was conducted between May and December 2017.

Survey procedure

The questionnaire and survey procedures were approved by the Human Research Ethics Committees of the Cancer Council Victoria, the NSW Population and Health Research Ethics Committee, and the NSW Department of Education and Communities. The survey was also endorsed by the Catholic Education Commission and the Association of Independent Schools of New South Wales.

Principals of selected schools were contacted by the NSW Ministry of Health's Centre for Epidemiology and Evidence to obtain permission to conduct the survey at their schools. If a school refused, they were replaced by the school nearest to them within the same sector. The aim was to survey 80 students from each participating school. For junior secondary, one class of 20 students (and 20 replacements) were randomly selected from each of Years 7-10; for senior secondary, two classes of 20 students (or 40 students and 40 replacements) were randomly selected from each of Years 11-12. A brochure and consent form was sent to the parents of each selected student and replacement. Consent forms were returned to the school and the school held the list of students who had parental consent. Written consent was sought from students with parental consent before the survey.

In 2017, McNair Ingenuity Research Pty Ltd was contracted to administer the pencil-and-paper questionnaire on the school premises. If a student from the sample list was not present at the time of the survey, a student from the replacement list for that year was surveyed. Students from different years were surveyed together. Students answered the questionnaire anonymously.

Survey instrument

The survey instrument was a written self-completion questionnaire, which included questions on alcohol, demographics, height and weight, injury, nutrition, physical activity, psychological distress, sedentary behaviour, substance use, sun protection (including sunburn experience and solarium use), and tobacco use.

Coding and data entry

Responses were coded and the data entered onto a database by the Centre for Behavioural Research in Cancer at The Cancer Council Victoria. After data entry, the data were cleaned and prepared for data analysis. Students whose questionnaires had a large amount of missing data or whose responses were extreme were removed from the dataset before analyses started. In the analysis, responses were excluded if the respondent gave contradictory or multiple responses or did not answer the question. However, these respondents remained in the analysis for the questions that they had validly completed. Cleaning of data relating to questions about the use of alcohol, tobacco, or other substances involved checking for inconsistencies in reported use across time periods (lifetime, year, month, and week). This cleaning procedure ensured maximum use of data and operated on the principle that the student's response about personal use in the most recent time period was accurate.

Data analyses and reporting

School students aged 12-17 years were included in the analysis. To ensure that disproportionate sampling of any school type, age level, and gender grouping, did not bias the prevalence estimates, data were weighted to bring the achieved sample into line with the population distribution. Reported prevalence estimates are based on these weighted data. Information about the enrolment details of male and female students in each age group at Government, Catholic and Independent schools was obtained from the Australian Bureau of Statistics (ABS Cat no. 4221.0, 2017)

Data were analysed using SAS version 9.4 (SAS Institute 2012). The SURVEYMEANS procedure in SAS was used to analyse the data and calculate point estimates and 95 per cent confidence intervals for the estimates. The SURVEYMEANS procedure calculates standard errors adjusted for the survey's design. It uses the Taylor expansion method to estimate sampling errors of estimators based on the stratified random sample (SAS Institute 2009). Estimates are presented for each response or indicator and by age group, sex, Local Health District (LHD) group and year where possible. Although figures are provided in every instance in the tables, if the estimates are not reliable because of small sample sizes the estimate is not shown in the graph. Where possible, indicators have been aligned with those collected previously, so that trends can be examined. Analysis of change over time is compared across two time periods, between the base survey year and current survey year, and between the previous survey year and the current survey year. The base survey year for particular indicators may vary, as the survey instrument has changed over time.

The 95 per cent confidence interval provides a range of values that should contain the actual value 95 per cent of the time. In general, a wider confidence interval reflects less certainty in the estimate for that indicator. The width of the confidence interval relates to the differing sample size for each indicator. A wider confidence interval reflects less certainty in the estimate. If confidence intervals do not overlap then the observed estimates are significantly different. If confidence intervals overlap slightly the observed estimates may be significantly different but further testing needs to be done to establish that significance. For a pairwise comparison of subgroup estimates, the p-value for a two-tailed test was calculated using the t-test for differences in means from independent samples, where appropriate (HealthStats NSW, 2015).

The Local Health District (LHD) was derived from the student's residential postcode. Although it was not possible to report on all indicators by LHD because of the survey's design, it was however possible to report by LHD groups by grouping some LHDs (i.e Central Coast and Northern Sydney; South Eastern Sydney, Sydney and Illawarra Shoalhaven; Western Sydney and Nepean Blue Mountains; Mid North Coast and Northern NSW; Murrumbidgee and Southern NSW; and Western NSW and Far West).

Characteristics of final sample in 2017

A total of 3,700 students in Years 7-12 were surveyed between May and December 2017. Just over half (52.7 per cent) were from Government schools, 24.4 per cent were from Catholic schools, and 22.9 per cent were from Independent schools. The final sample's sex distribution was 47.2 per cent male and 52.8 per cent female and the age distribution was 57.2 per cent aged 12 to 15 years and 42.8 per cent were aged 16-17 years. When the sample were weighted to the secondary school student population in NSW by age and sex, 51.0 per cent were male and 49.0 per cent were female, 69.8 per cent were aged 12-15 years and 30.2 per cent were aged 16-17 years. Information about the enrolment details of male and female students in each age group at Government, Catholic and Independent schools was obtained from the Australian Bureau of Statistics (ABS Cat no. 4221.0).

In 2017, the sample also consisted of 6.3 per cent Aboriginal or Torres Strait Islander students, which was similar to the NSW proportion of Aboriginal or Torres Strait Islander students in 2017 of 5.8 per cent (ABS Cat no. 4221.0). The main language spoken at home in the final sample was English (72.5 per cent), followed by English and another language (24.0 per cent), and another language only (3.4 per cent).

Important caveats and data limitations

As this Survey was only answered by a sample of secondary students in New South Wales, it is important to note that estimates of health behaviours are subject to a margin of error.

Furthermore, self-reports of certain health behaviours are known to be subject to social desirability bias, which is a term used to describe the tendency for people to present a favourable image of themselves when responding to surveys. This may lead to the prevalence of certain positive behaviours being overstated, with undesirable or negative behaviours being understated. While an anonymous self-complete questionnaire, such as that used for this survey, provides respondents with the greatest level of privacy when responding to sensitive questions, it is possible that certain health behaviours may be under or over-estimated in this report.

References

Australian Bureau of Statistics. Schools, Australia 2017. Catalogue no. 4221.0. Canberra: ABS, 2017. Available at http://www.abs.gov.au/ausstats/abs@.nsf/mf/4221.0

Centre for Epidemiology and Evidence. NSW School Students Health Behaviours Survey Methods. CEE, NSW Ministry of Health. Available at http://www.health.nsw.gov.au/surveys/student/Pages/default.aspx

Centre for Epidemiology and Evidence, NSW Ministry of Health. Confidence intervals method paper. Available at http://www.healthstats.nsw.gov.au/Resources/Confidence_Intervals.pdf

SAS Institute 2012. The SAS version for Windows version 9.4. Cary, NC: SAS Institute Inc., 2012.