The NATSIHS is conducted in all states and territories and includes remoteand non-remote areas. The 2004-05 sample was 10,000 persons/5200 households, with a response rate of 81 per cent of households. The NHS is conducted in all states and territories excluding very remote areas. This exclusion has a small impact only on national and state estimates, except for the Northern Territory, where such persons comprise over 20 per cent of the population. As a consequence of this exclusion, comparisons between Indigenous and non-Indigenous people in remote areas are not available. The 2004-05 NHS response rate was 91 per cent of households. Both surveys are weighted to account for non-response. Since it is derived from sample surveys, this indicator is subject to sampling error which occurs because only a small proportion of the population is used to produce estimates that represent the whole population. Sampling error can be reliably estimated and is based on the statistical methods used to design surveys. Overall, this indicator has an relative standard error (RSE) of less than 25 per cent for all states and territories, except for Indigenous estimates of ‘high level of physical activity’ for Victoria (27 per cent), South Australia (26 per cent) and the Northern Territory (40 per cent) which should be used with caution. Finer levels of disaggregation (e.g. by the inclusion of other cross classifying variables) may result in higher levels of sampling error. |