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

ESTIMATION

1 PES estimation involves assigning a weight to each selected PES dwelling and then to each person for whom a PES response was obtained. The weights attached to PES persons allow the PES sample to represent the whole population of interest; i.e. all usual residents in Australia on Census night, including people in non-private dwellings (e.g. hotels, hospitals and jails) which are not covered by the PES dwelling sample.

DWELLING WEIGHTING

2 Dwelling weighting for the 2016 PES comprised two stages. For private dwellings selected in the PES that were found in the Census, the first stage of weighting adjusted the PES selection weight (the inverse of the probability of a dwelling being selected in the PES sample) such that the adjusted weights added up to the Census private dwelling count within categories based on geography and dwelling characteristics. A first-stage weight adjustment was also applied to private dwellings selected in PES that were missed in the Census. For dwellings in the Discrete Community sample, a similar first-stage weight adjustment was applied based on dwelling counts for communities within each state and territory.

3 The second stage of dwelling weighting applied a non-response adjustment so that the responding PES dwellings represented other dwellings from which no response was obtained.

PERSON WEIGHTING

4 The initial stage of person weighting adjusted the dwelling weights to ensure that the PES estimates of people counted (in the Census) in private dwellings and Discrete Community dwellings (other than the non-contact sector) in a set of benchmark categories matched the actual Census counts for these categories. The weight adjustment applied to a person did not depend on whether they responded in the Census, but only on characteristics of the person as reported in the PES.

5 As a final step in weight adjustment, the initial person weights were adjusted so that the PES estimates also represented people in non-private dwellings, such as hotels, hospitals and jails, which were not covered by the PES. The person weighting step in PES processing calculates weights for all PES records, including those relating to the non-contact sector.

6 Intuitively, a good set of weights for the PES should ensure that if the PES were used to estimate the actual Census count, the PES would get the right answer. The above step ensured this was the case. Technically, this is a desirable property for a set of PES weights to have since there is a very strong relationship between the actual Census count and the count that the Census should have made.

7 In 2016, the ABS again used the Prediction Regression (PREG) estimator, which was developed and used in 2006 and 2011. This method was introduced in 2006 to account for overlapping benchmark categories and the situation where people gave different responses between PES and Census. The resulting estimates are categorised by the PES response (i.e. where they should have been counted), rather than the Census response. A detailed description of the PREG estimator can be found in Chipperfield et al. (2016)[1].

ESTIMATES OF THE POPULATION AND NET UNDERCOUNT

8 The PES population estimate for a category of person (e.g. males) is equal to the sum of person-weights of those persons who should have been counted in that category in the Census. The final PES estimate is then adjusted for consistency with Census counts, by adding on the persons who were counted in that category in responding Census dwellings (i.e. dwellings from the contact sector) and subtracting the weighted estimate of these people.

9 Net undercount for any category of person is then the difference between the final PES population estimate for that category and the actual Census count (including imputed persons in non-responding dwellings). This calculation takes into account the components described in Components of Net Undercount on the Summary tab.

ENDNOTES

1 Chipperfield J, Brown J and Bell P 2016. ‘Estimating the Count Error in the Australian Census’, Journal of Official Statistics, vol. 33, pp. 1–17.