Pandemic insights into Australian smokers, 2020-21

Smoking estimates in this release are drawn from the Smoker Status, Australia dataset which is an experimental dataset built from common content included in all surveys conducted from July 2020 to June 2021.

The household surveys included in this experimental dataset are:

• National Health Survey (NHS) 2020-21
• General Social Survey (GSS) 2020
• Survey of Income and Housing (SIH) 2020-21
• Time Use Survey (TUS) 2020-21
• National Study of Mental Health and Wellbeing (NSMHW) 2020-21.

These surveys collected a standard set of information which were pooled to produce the Smoker Status dataset, including age, sex, country of birth, main language, employment, education, migrant and visa status, and current smoker status.

Pooling this content from multiple surveys has produced a large sample size, for the purpose of providing more accurate national smoking estimates compared to any individual survey. Pooled smoking data will be released annually for 2020-21, 2021-22 and 2022-23 financial years.

The surveys used to create this dataset were collected during the COVID-19 pandemic. To maintain the safety of survey respondents and ABS Interviewers, surveys were primarily collected via online, self-complete forms with some telephone and face-to-face interviews conducted where possible. Non-response is usually reduced through Interviewer follow up of households who have not responded. As this was not possible during lockdown periods, there have been significant impacts on response rates and sample representativeness. Due to these changes, comparisons to previous smoking data over time are not recommended.

Smoking data has previously been pooled from the 2017-18 National Health Survey (NHS) and the 2017-18 Survey of Income and Housing (SIH) and the smoking estimates were published in December 2018.

While similar in that both pooled datasets have been used to produce smoking estimates, the 2020-21 Smoker Status dataset has expanded data sources, collection methodologies and content. It should not be used to create a time series with 2017-18 (and previous) data for smoking trends. The 2020-21 Smoker Status dataset is considered a break in series, and reflects the specific time point only.

The scope of the dataset is as follows:

• Usual residents (URs) in Australia aged 15 years and over living in private dwellings were in scope for all household surveys, noting that the SHWB had an age scope of 16 to 85 years
• Both urban and remote areas in all states and territories, except for very remote parts of Australia and discrete Aboriginal and Torres Strait Islander communities
• Members of the Australian permanent defence forces living in private dwellings and any overseas visitors who have been working or studying in Australia for the last 12 months or more, or intend to do so.

The following people were excluded:

• Visitors to private dwellings
• Overseas visitors who have not been working or studying in Australia for 12 months or more, or do not intend to do so
• Members of non-Australian defence forces stationed in Australia and their dependents
• Non-Australian diplomats, diplomatic staff and members of their households
• People who usually live in non-private dwellings, such as hotels, motels, hostels, hospitals, nursing homes and short-stay caravan park (people in long-stay caravan parks, manufactured home estates and marinas are in scope)
• People in very remote areas
• Discrete Aboriginal and Torres Strait Islander communities.

Households were randomly selected to participate in the surveys used for the pooled dataset. If the randomly selected person was aged 15-17 years, parental consent was sought for the interview to proceed.

The total sample pooled from the five surveys was 30,564 households and 42,117 persons.

The mode of collection varied across surveys due to the timing of enumeration in relation to the COVID-19 pandemic and specific survey requirements. Modes of collection utilised were:

• Self-completed online form
• Telephone interview with an ABS trained Interviewer
• Face-to-face interview with an ABS trained Interviewer (limited numbers)
• Self-completed paper form (limited numbers).

All household surveys collected a common set of information including:

• Demographics - Age, Sex, Country of Birth, Main language spoken, Marital status
• Household details - Type, Size, Household composition, Tenure, SEIFA, Geography
• Labour force status
• Educational attainment
• Self-assessed health status
• Migrant and Visa status
• Current smoker status.

See the data item list for the Smoker Status, Australia dataset for more details about content.

As only a sample of people in Australia were surveyed, their results needed to be converted into estimates for the whole population. This was done through a process called weighting:

• Each person or household is given a number (known as a weight) to reflect how many people or households they represent in the whole population
• A person or household’s initial weight is based on their probability of being selected in the sample. For example, if the probability of being selected in the survey was one in 45, then the person would have an initial weight of 45 (that is, they would represent 45 people).

The person and household level weights are then calibrated to align with independent estimates of the in-scope population, referred to as ‘benchmarks’. The benchmarks use additional information about the population to ensure that:

• People or households in the sample represent people or households that are similar to them
• The survey estimates reflect the distribution of the whole population, not the sample.

Benchmarks align to the estimated resident population (ERP) at December 2020, aged 15 years and over, which was 9,782,954 households and 20,285,817 people (after exclusion of people living in non-private dwellings, very remote areas of Australia and discrete Aboriginal and Torres Strait Islander communities).

There was no imputation for missing data on the pooled dataset. Any records with an unacceptable level of missing data were removed. However, if the level of missing data was minimal, the records were kept with ‘not stated’ values where needed.

Pooled sample counts and weighted estimates are presented in the table below.

Two types of error are possible in estimates based on a sample survey: 

• Non-sampling error    
• Sampling error.

Non-sampling error is caused by factors other than those related to sample selection.  It is any factor that results in the data values not accurately reflecting the true value of the population.

It can occur at any stage throughout the survey process. Examples include:

• Selected people that do not respond (e.g. refusals, non-contact)
• Questions being misunderstood 
• Responses being incorrectly recorded 
• Errors in coding or processing the survey data.

Sampling error is the expected difference that can occur between the published estimates and the value that would have been produced if the whole population had been surveyed. Sampling error is the result of random variation and can be estimated using measures of variance in the data.

One measure of sampling error is the standard error (SE). There are about two chances in three that an estimate will differ by less than one SE from the figure that would have been obtained if the whole population had been included. There are about 19 chances in 20 that an estimate will differ by less than two SEs.

A confidence interval expresses the sampling error as a range in which the population value is expected to lie at a given level of confidence. A confidence interval is calculated by taking the estimate plus or minus the MOE of that estimate. In other terms, the 95% confidence interval is the estimate +/- MOE. 

This release presents national smoking estimates for 2020-21. Commentary presents analysis by age groups, sex and selected population characteristics.

Data Cubes (spreadsheets) in this release contain smoking estimates, proportions and their associated measures of error.

Detailed microdata is also available on DataLab for users who want to undertake interactive (real time) complex analysis of microdata in the secure ABS environment.

The Census and Statistics Act 1905 authorises the ABS to collect statistical information, and requires that information is not published in a way that could identify a particular person or organisation. The ABS must make sure that information about individual respondents cannot be derived from published data.

To minimise the risk of identifying individuals in aggregate statistics, a technique called perturbation is used to randomly adjust cell values. Perturbation involves small random adjustment of the statistics which have a negligible impact on the underlying pattern. This is considered the most satisfactory technique for avoiding the release of identifiable data while maximising the range of information that can be released. After perturbation, a given published cell value will be consistent across all tables. However, adding up cell values in Data Cubes to derive a total may give a slightly different result to the published totals. The introduction of perturbation in publications ensures that these statistics are consistent with statistics released via services such as TableBuilder.