Research Paper no. 26 2007–08
Postcodes in electoral divisions (2006 electoral boundaries)
Statistics and Mapping Section
9 April 2008
This research paper provides a concordance of postcodes to Commonwealth electoral divisions, based on Australian Electoral Commission electoral roll enrolments. The concordance allows members, senators and other interested parties to convert administrative data collected by postcodes into electoral division totals.
Table 1 Electoral divisions by postcode
Table 2 Postcodes by electoral division
This paper provides a concordance of postcodes to Commonwealth electoral divisions based on enrolment data held by the Australian Electoral Commission. It is an update of Research Paper No. 11, 2003 04, titled Postcodes in electoral divisions (2003 electoral divisions) and it reflects both changes to postcodes since 2001 and changes to electoral division boundaries brought about by the 2005 redistribution in the Australian Capital Territory and the 2006 redistributions in New South Wales and Queensland.
The concordance will enable members, senators and other interested parties to convert administrative postcode data, published by organisations such as Centrelink and the Australian Taxation Office, into electoral division totals.
The concordance was generated by processing electoral roll data supplied by the Australian Electoral Commission (AEC). These data are a snapshot of all enrolments on the electoral roll as at 31 December 2007. The data do not identify individuals but are simply a count of enrolments classified by locality, postcode of residence and electoral division. The AEC devotes significant resources to ensuring the completeness and accuracy of enrolments as well as validating address information and allocating each address to the correct electoral division. As a result, the electoral roll is considered to be the most accurate source of data containing counts of persons linked by postcode and electoral division, although it should be noted that it only covers persons aged 18 years and over.
The previous concordance linking postcodes and 2003 electoral divisions, created by the Parliamentary Library several years ago, was based on 2001 Census data at the Census Collection District (CCD) level. The 2006 Census could have been used to update this concordance, however CCDs do not align with postcode boundaries and each CCD is simply allocated to a postcode of best fit. The Australian Bureau of Statistics refer to these census derived boundaries as Postal Areas to distinguish that they are only an approximation of the actual postcodes and some Australia Post postcodes are not even included in the Postal Area classification. For this reason it was decided that a concordance based on electoral roll data would be more accurate than a concordance based on CCD data. This approach is confirmed by validation checks described below.
The initial concordance created from AEC enrolment data was validated and enhanced as a result of several edit checks. The first validation check was to compare the list of postcodes based on enrolment data against the list of postcodes based on 2006 Census data. The enrolment data file contains around 130 additional postcodes that are not in 2006 Census data files, thus the enrolment file has a more complete coverage of postcodes. There are also 14 additional postcodes that appear in the 2006 census data files that are missing from the enrolment file. Two of these postcodes (6435 and 6723) do not appear to be valid postcodes based on current Australia Post information and they were ignored. The remaining 12 postcodes mainly relate to university campuses and these postcodes were added to the final concordance to increase its coverage.
The second validation check was to compare the postcode weights based on enrolment data against postcode weights based on 2006 Census data. Significant differences of more than 15 per cent in a postcode weight were examined together with cases where a postcode is missing from one concordance. These differences were examined by looking at a spatial map of postcode boundaries and electorate boundaries. In many cases a visual examination of the spatial boundaries will not allow a definitive assessment of which weight is better. For example, if a postcode has a sizeable overlap with two different electorates then a visual examination of a map showing the postcode and electorate boundaries will not allow an assessment of whether a weight of 40 per cent or 60 per cent is more reasonable. However there were a number of cases where the postcode weights based on enrolment data can be assessed as being superior to postcode weights based on 2006 Census data. These cases relate to situations where a visual examination of a postcode shows that it partially spans an electorate but there is not a 2006 Census based postcode weight. An example of this situation is the postcode 3446 that has a weight of 77.7 per cent in Bendigo and 22.3 per cent in Ballarat, based on enrolment data, whereas the 2006 Census data allocates all of the postcode to the Bendigo electorate. Another example is the postcode 5356 that has a weight of 89.5 per cent in Barker, 8.5 per cent in Wakefield and 2.9 per cent in Grey, based on enrolment data, whereas 2006 Census data allocates all of the postcode to Barker. This validation check supports the use of an enrolment based concordance in preference to a CCD based concordance.
To increase the coverage of the concordance based on enrolment data, the list of postcodes was matched against both the previous concordance based on 2001 Census data and the list of postcodes for which there are current spatial boundaries. The additional postcodes found from this matching process that are current postcodes (based on Australia Post information) were added to the concordance. Based on these edit checks, postcodes 3671, 6038, 6452, and 6712 were found to have a current spatial boundary and were added to the concordance. The postcode 6733 from the previous concordance was also added to the concordance.
A final validation check on the concordance examined small cells where a concordance record was based on a very small number of enrolments and in some cases the concordance record was omitted.
The concordance information in this paper shows which postcodes are assigned to which electoral division and the percentage of the postcode (the weight) that is attributable to the electoral division. The information is shown in two tables: The first table is sorted by postcode whereas the second table is sorted by electoral division name. Both tables contain the same data.
There are 2,643 separate postcodes covered by the concordance and 2,037 of these postcodes lie wholly within an electorate. Users should note that some postcodes span state/territory boundaries. Also there are many non-residential postcodes which are not able to be assigned to an electoral division of residence so they are not in the scope of the concordance. Examples of a non-residential postcode includes 2001 which is a post office box and 2123 which is a special postcode used for large volumes of mail.