• Summary
• Downloads
• Explanatory Notes
• Related Information
• Past Releases

Introduction

4.1. The accuracy of population estimates is dependent on the quality of population census data, the estimates of census net undercount derived from the census Post-Enumeration Survey and the statistics of the components of population change which are used to update the census year population. This chapter discusses these various sources of data and their accuracy.


Census count of residents

4.2. The base population for population estimates, which is derived from the latest available Population Census, comprises residents only, (ie. people whose place of usual residence is in Australia). The Census of Population and Housing in Australia, however, is currently enumerated according to their actual location and aims to count everyone who spends Census night in Australia except for foreign diplomats and their families. This includes Australian residents in Antarctica and people in the territories of Jervis Bay, Cocos (Keeling) Islands and Christmas Island. The other Australian territories of Norfolk Island and minor islands such as Heard and McDonald Islands, are outside the scope of the Australian Census. Australians out of the country on Census night are generally out of scope of the Census.

4.3. The Census includes people on vessels in or between Australian ports, on board long-distance trains, buses or aircraft and on oil or gas rigs off the Australian coast. People entering Australia before midnight on Census night are counted while people leaving an Australian port for an overseas destination before midnight on Census night are not. Visitors to Australia are included regardless of how long they have been in the country or how long they plan to stay. However, the census question on address (see paragraph 2.6; also Usual residence) provides the information needed to identify these non-residents so that they can be excluded for the purpose of estimating the population.

4.4. The same census question provides the information needed to identify the residents of each State. For example, the residents of a particular State are only those people whose place of usual residence, as declared in the question, is in that State.

4.5. Similarly, the same principle applies for smaller areas such as Statistical Local Areas.

Non-response to census questions

4.6. Besides the question on address, questions on age, sex, marital status, country of birth and Indigenous origin are particularly important in the context of estimating population. Instances in which people do not answer these questions therefore have a bearing on the accuracy of census counts, although information is imputed by the ABS from other details supplied on the census form or, failing that, using a variety of imputation techniques. Details are imputed for all instances in which there is no response to questions on age and sex.

4.7. Table 4.1 shows that non-response to the questions which are of importance to population estimates has remained low in recent censuses. There is, however, a rising trend in non-response to the country of birth question, though from a low base.

	4.1: Rate of Non-response to Census Questions, 1976 to 1996 (per cent)
	Variable	Census
		1976	1981	1986	1991	1996
	Sex	(a)	(a)	(a)	(a)	(a)
	Age	2.8	1.8	1.8	2.0	2.2
	Usual residence	1.5	1.3	1.1	(c)	2.5
	Marital status (b)	4.1	3.6	3.3	6.9	2.0
	Country of birth	(c)	1.2	1.6	2.2	3.5
	Indigenous origin	(c)	3.8	1.7	3.2	3.0
	(a) Not available but believed to be insignificant. (b) As a per cent of the population of all ages. However a considerable proportion of non-responding persons were aged under 15 years and never married. All persons aged under 15 were coded to never married (c) Not available.