Australian secondary school students' use of tobacco, alcohol, and over-the-counter and illicit substances in 2011

# 2.6 Data analyses

Analyses cover school students aged 12 to 17 years. To ensure that disproportionate sampling of any state/territory, education sector, age and gender grouping did not bias the prevalence estimates, data were weighted to bring the achieved sample into line with the distribution of the population of 12 to 17-year-olds in secondary schools throughout Australia. The prevalence estimates reported here are based on these weighted data. Enrolment details of male and female students in each age group at government, Catholic and independent schools were obtained from the Australian Bureau of Statistics (refer to Appendix 3). Respondents were not included in the analyses of particular questions if they gave contradictory or multiple responses, or did not answer the question. However, these respondents were included in the analysis of other questions if these had been validly completed.

As this report is based on data from a sample and not on a census of the total population, it is necessary to allow for sampling error. Prevalence estimates are provided for information, regardless of their levels of statistical reliability. For percentages or proportions, the sampling error is generally indicated by the 95% confidence interval. The 95% confidence interval is based on both the number of students in the specific group examined (e.g.12-year-old males) and the percentage reported (e.g. 15%). The confidence interval is larger when the sample size is small and the estimate is around 50%. For the 2011 survey, the largest confidence interval will be found for 12-year-old males, as this group has the smallest sample size (n=887). The 95% confidence interval for 12-year-old males around an estimate of 50% is 3.3%, meaning that the actual percentage will be between 46.7% and 53.3%. Thus, using 95% confidence intervals, the estimates of the prevalence of use for the different substances among different age and gender groups reported here are within 3.3% or better of the true population values. When interpreting results, readers should refer to the table of 95% confidence intervals associated with the sample size achieved for each age and gender group (refer to Appendix 4). The reader should be aware that some results, especially those for most illicit substances, may be statistically indistinguishable from zero based on the 95% confidence interval associated with an estimate.

Given the large overall sample size and in accordance with previous practice, only those results associated with a p value of <0.01 were taken to be statistically significant.

In addition, caution should be used when interpreting the reported findings as they are based on self-reported data.