National Healthcare Agreement: P64b-Indigenous Australians in the health workforce, 2010 QS
Identifying and definitional attributes
|Metadata item type:||Quality Statement|
|Registration status:||Health, Retired 12/03/2015|
|Indicators linked to this Quality statement:|
National Healthcare Agreement: P64b-Indigenous Australians in the health workforce, 2010 Health, Superseded 08/06/2011
|Institutional environment:||The Census is conducted by the Australian Bureau of Statistics (ABS) under the authority of the Census and Statistics Act 1905. The Act requires that a Census be conducted every five years, and at other times as prescribed. The Census includes all people in Australia on Census night, with the exception of foreign diplomats and their families.
The Census is collected, processed and published within a legislative framework that ensures the independence and impartiality from political influence of the ABS, and the confidentiality of respondents.
All ABS officers (including temporary employees) are legally bound to secrecy under the Act never to release identifiable personal information to any person or organisation outside the ABS. Section 19 of the Act makes it an offence for any past or present ABS officer to divulge, either directly or indirectly, any confidential information collected under this Act.
For more information on the institutional environment of the ABS, including the legislative obligations of the ABS, financing and governance arrangements, and mechanisms for scrutiny of ABS operations, please see ABS Institutional Environment.
|Timeliness:||The Census is conducted every five years, with the most recent Census conducted in 2006. Census data on Occupation are available approximately 14 months after the Census date.|
|Accessibility:||See Census Products for standard products available from the Census. Data is also available on request.|
|Interpretability:||For information to aid interpretation of Census data, see Census Reference and Information on the ABS website, www.abs.gov.au.
For more information comparing data sources of Indigenous health labour force statistics, see the Australian Institute of Health and Welfare (AIHW) publication Aboriginal and Torres Islander health labour force statistics and data quality assessment.
|Relevance:||The Census collects information on Occupation for all employed people aged 15 years and over. Two questions are used in the Census: ‘In the main job held last week, what was the person’s occupation? – Give full title’; and ‘What are the main tasks that the person usually performs in the occupation?’ Occupation is then coded according to the Australian and New Zealand Standard Classification of Occupations (ANZSCO) using these written responses. |
The 'Commonwealth working definition' of an Aboriginal or Torres Strait Islander is ‘a person of Aboriginal or Torres Strait Islander descent who identifies as an Aboriginal or Torres Strait Islander and is accepted as such by the community in which he or she lives’. In statistical collections, it is not feasible to collect information on the community acceptance part of this definition, and therefore questions on Indigenous Status relate to descent and self-identification only. In practice, people are asked if they are of Aboriginal or Torres Strait Islander origin. Those who identify themselves as of Aboriginal and/or Torres Strait Islander origin are classified as Indigenous persons.
|Accuracy:||Extensive effort is put into Census form design, collection procedures and processing to minimise potential error. There are four principle sources of error in Census data: respondent error, processing error, partial/nonresponse and undercount.
Information on Indigenous status is captured automatically from check box responses so the risk of processing error is minimal. Sample checks of the data are undertaken to ensure an acceptable level of quality. The overall non-response rate for Indigenous status for 2006 was 5.7%. Most of this non-response (4.1%) is attributable to persons in dwellings which were occupied on Census Night but which did not return a completed form (i.e. dwelling non-response). Persons are imputed into these dwellings (based on information on the number of persons in the dwelling if available), along with demographic characteristics, however Indigenous status is not imputed and remains ‘not stated’. In some states and the Northern Territory, the number of Aboriginal people counted in the Census in some urban areas and regional towns is below what might be expected based on 2001 Census results. Affected areas appear mostly to have been areas where there were issues with the recruitment and retention of Census Collectors. These areas tend to have higher numbers of non-responding dwellings, and evidence suggests that these dwellings may have contained higher than average proportions of persons of Indigenous origin. It may therefore be the case that Indigenous persons in the health workforce may be more likely to not respond to the Census than non-Indigenous persons in the health workforce, which would lead to an underestimation of this indicator value.
Undercounting of Indigenous Australians may also introduce an underestimation bias into the results. The net Census undercount for Indigenous Australians was estimated at 11.5%, higher than the net undercount for non-Indigenous Australians (2.7%).
Occupation is coded with reference to the responses to the two occupation questions as provided by respondents. The quality of information on Occupation is therefore affected by the level of detail provided by respondents and the ease with which responses can be coded.
A principle of occupation coding is to allocate responses to the most descriptive and detailed level possible from the information supplied by respondents. If a response is not detailed enough to allow coding to the most detailed level of the classification, it is coded to the most detailed level possible, and a not further defined (nfd) code is allocated. Standard automatic coding procedures were used to code 62.4% of responses, with remaining responses coded using clerical procedures. All processes are subject to sample checks to ensure an acceptable level of quality.
Census data can be used for the analysis of population characteristics at fine geographic levels and for small sub-groups. However at detailed levels of disaggregation, outliers (unusual results) may become more apparent. Quality statements for the data items are available, (see Census Data Quality Statement, and data quality statements for Occupation and Indigenous Status).
|Coherence:||The data items used to construct the measures are consistent and comparable, and support assessment of change between Census years.
The Australian and New Zealand Standard Classification of Occupation (ANZSCO) has been used in all ABS collections with education items since 2006 and allows the occupation items between different collections to be compared. Census responses were also coded to the previous version of the classification, ASCO, to facilitate comparison with previous Censuses.
There are three main sources of data on the health workforce. The five yearly Census of Population and Housing, the monthly ABS Labour Force survey, and AIHW Health Labour Force surveys.
As a sample survey, the ABS Labour Force survey is subject to sampling error. While sampling error will generally be within acceptable levels for national and state/territory level information, detailed cross-tabulations (such as Indigenous persons in particular occupations by state) will be subject to high levels of sampling error and be subject to confidentiality concerns. For this reason, the monthly ABS Labour Force survey was not considered as a potential data source for this indicator.
There are a number of differences between the Census and the AIHW Labour Force surveys that might affect their comparability. Reference should be made to the Data Quality Statement for Indicator 64a when comparing the two data sources.
The AIHW surveys use a sample frame of persons registered with the relevant registration boards for that profession, regardless of employment status. Respondents are generally targeted via the registration renewal process for their profession in each jurisdiction and the entire population of registrants is sent a questionnaire. Information is then collected in the survey on labour force status, demographic characteristics (including Indigenous status using the standard identification question), specialist field, and qualifications. The AIHW surveys are voluntary and have relatively high rates of nonresponse. Of particular relevance to the quality of estimates for Indigenous Australians, the response rate in 2007 in the Northern Territory was approximately 27%.
To account for non-response to the AIHW surveys, estimates are weighted or benchmarked to registration numbers provided by state and territory registration boards. This relies on up to date and reliable registration data, which may differ across time and between jurisdictions.
Some Indigenous health workers are likely to be missed through not identifying as Indigenous on the survey form. In addition, among those who do not respond to the survey, some will be Indigenous persons. If the characteristics of those that respond are different from those that do not respond, bias in the distribution of characteristics (including Indigenous status) will be introduced with this non-response adjustment. However it is not known whether the final survey estimates were overestimates or underestimates.
Differences in published numbers may also be as a result of differences between the Census occupation definitions according to ANZSCO, and the definitions used in the AIHW surveys.
The number of Indigenous health workers may change over time due to real changes in the number of Indigenous health professionals, changes in response rates, or changes in Indigenous identification. The number of persons with a characteristic seldom changes rapidly. A good indication of quality of a particular data source is therefore its stability over time. Change in variables over time should be gradual. Large fluctuations over time are indicative of a data quality issue and/or small numbers involved.