2940.0 - Census of Population and Housing: Details of Overcount and Undercount, Australia, 2016 Quality Declaration 
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 27/06/2017   
   Page tools: Print Print Page Print all pages in this productPrint All

CENSUS COUNTS

USUAL RESIDENCE

1 The Australian Census counts people where they stayed on Census night. This means the Census was conducted on an actual location or place of enumeration basis.

2 For usual residents of Australia, place of usual residence for the 2016 Census was defined as the address at which a person has lived, or intends to live, for six months or more in 2016. While for most people their usual residence was the same as their actual location on Census night, some people spent Census night at a place other than where they usually lived. Thus, their place of enumeration and their place of usual residence were different.

3 People visiting Australia on Census night were included in the Census counts on a place of enumeration basis but not those on a place of usual residence basis.

4 Usual residents of Australia who were temporarily overseas on Census night were not included in Census counts on either a place of usual residence or place of enumeration basis. Counts of these people were accounted for, however, in the Estimated Resident Population (ERP) of Australia. For information on the calculation of ERP, see the ABS publication Australian Demographic Statistics, December 2016 (cat. no. 3101.0), released on 27 June 2017.

5 Census counts presented in this publication are on a place of usual residence basis (unless otherwise noted).

IMPUTATION

6 In addition to complete person records, Census counts include:

  • persons whose Census form was partially completed
  • imputed person records for non-responding dwellings.

7 During Census processing, values for Age, Sex, Registered marital status, and State/Territory of usual residence were imputed if they were left blank on the Census form (i.e. missing item response from an otherwise responding person). These variables are needed for population estimates, so they were imputed using other information on the Census form. These were also the only variables imputed for the imputed persons in non-responding dwellings.

8 Values for all other variables left blank (e.g. Indigenous status, Country of birth or Occupation) were set to not stated or not applicable, depending on the (possibly imputed) age of the person.

9 While the method of imputation remained the same as 2011, the 2016 Census experienced a reduction in the amount of quantitative information they could use in their imputation model. Specifically, the ‘credible source’ information collected for non-responding dwellings (where possible) in previous Censuses was no longer collected. This meant that information, such as whether the dwelling was occupied on Census night and the number of males and females in that dwelling, was no longer available to guide the imputation process. Instead, imputation was based solely on the values observed for responding persons.

10 For further information about imputed persons and how they are treated in PES, see Components of Net Undercount on the Summary tab.

RESOLUTION OF CENSUS NOT STATED VALUES FOR PES

11 The PES uses Census data items to form benchmark categories for weighting and estimation purposes, including Indigenous status and Country of birth. In cases where one or both of these two items had been set to not stated in the Census, a value was imputed during PES processing so they could be used for benchmarks in PES estimation.

12 The imputation method involved imputing both variables together. For benchmarking purposes, 12 non-overlapping categories were defined (e.g. Australian born and not of Aboriginal or Torres Strait Islander origin, or born in one of the ten highest ranked countries and not of Aboriginal or Torres Strait Islander origin). Imputation was performed within a number of imputation classes based on geography (SA2), the Census form type, age and sex. The imputed value was based on the proportion of stated values of respondents in the same imputation class. This method also required the assumption that anyone who was imputed as Aboriginal or Torres Strait Islander was also born in Australia.

CENSUS DWELLING TYPES

13 Although all persons resident in Australia should have been counted in the Census, not all dwellings would have received a Census form. This is because not all dwellings were habitable or they were missing from the ABS Address Register, or, in the case of a diplomatic dwelling, did not contain people within scope of the Census.

14 Census defines private dwellings as structures established for self-contained accommodation. A private dwelling is most often a house or flat. It can also be a caravan, houseboat, tent, or a house attached to an office, or rooms above a shop. Private dwellings are either occupied or unoccupied:
  • Occupied – private dwellings that were occupied by one or more persons on Census night
  • Unoccupied – private dwellings which are habitable but were not occupied on Census night. This includes unoccupied holiday houses, vacant houses, and unoccupied apartments. Unoccupied dwellings in caravan/residential parks, camping grounds, marinas, and manufactured home estates were not counted in the Census, with the exception of unoccupied residences of managers/caretakers of such establishments.

15 Private dwellings may also be classified as non-responding or a refusal in the Census:
  • Non-responding – private dwellings where the Census Field Officer was unable to make contact with a householder within five visits, and was unable to verify that the dwelling was unoccupied on Census night. This excludes dwellings where the householder contacted the ABS to advise that the dwelling was unoccupied
  • Refusal private dwellings where the householder refused to participate in the Census.

16 For PES purposes, dwellings that provided a Census form after the commencement of PES enumeration were flagged as late returns for PES estimation. For information on the treatment of late returns, see Components of Net Undercount on the Summary tab.

CENSUS DATA QUALITY

17 Census data are subject to a number of inaccuracies resulting from errors by respondents or mistakes in collection or processing. While many of these are corrected by careful processing procedures, some remain. The effect of the remaining errors is generally slight, although it may be more important for smaller groups in the population.

18 The main kinds of errors occurring in the Census are:
  • Partial non-response – When completing their Census form, some people do not answer all questions which apply to them. When this occurs, some items might be imputed while others are set to not stated or not applicable, as discussed in the section on Imputation, above
  • Processing error Much of the Census data is recorded using automatic processes, such as scanning, Intelligent Character Recognition and other automatic processes. Quality assurance procedures are used during Census processing to ensure processing errors are kept to an acceptable level. Sample checking is undertaken during coding operations and corrections are made where necessary
  • Respondent error The Census is self-enumerated, meaning householders are required to complete the Census form themselves, rather than having the help of a trained interviewer, and may be completed by one household member on behalf of others. Errors can be introduced if the respondent does not understand the question, or does not know the correct information about other household members. Self-enumeration carries the risk that wrong answers could be given, either intentionally or unintentionally.

19 For further information on sources of error in the Census, see the ABS publication Census of Population and Housing: Census Dictionary, 2016 (cat. no. 2901.0), released on 23 August 2016.