2975.0.55.012 - Census Working Paper 96/4 - Fact Sheet 12 - 1996 Census Postal Code Information, 1996  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 15/09/1999   
   Page tools: Print Print Page Print all pages in this productPrint All

1996 CENSUS OF POPULATION AND HOUSING
FACT SHEET 12
1996 CENSUS POSTAL CODE INFORMATION


There are TWO different bases for tabulating 1996 Census statistics using Postcode as a reference. The ABS only produces ONE type of 1996 Census Digital Boundary which uses Postcode as the basis of the geographic classification, that is Postal Area (POA).


TABULATING CENSUS STATISTICS

Census data can be tabulated using two different methods:

        • Postal Area of Enumeration; and
        • Postcode of Usual Residence.
Postal Area of Enumeration (POA)

Postal Area (POA) is formed by allocating Collection Districts (CDs) to Australia Post postcode areas on a 'best fit' basis. (See P.4 Statistical Geography Volume 2 - Census Geographic Areas (Cat. No. 2905.0)) It is similar in concept to CD Derived Postcode in the 1991 census. The Postal Area classification primarily relates to place of enumeration data produced from the Population Census. That is individuals are tabulated in the CD in which they were counted (whether they were residents or visitors) and are included in the POA to which that CD is allocated. For example, if a person was enumerated in CD 1400701 then they would be included in POA 2016 because CD 1400701 has been allocated to this POA code.

Some Australia Post postcodes are not included in the Postal Area classification. In some cases a CD cannot be allocated to an Australia Post postcodes because of the application of the 'best fit' principle and as such these postcodes can not be included in the POA classification. This occurs in two cases:

        • where a CD covers 2 or more whole postcodes and, because of the classification rules, only one POA can be allocated to the CD, or;
        • where more than one CD partially covers a Postcode but all the CDs are allocated to other Postcodes which they also share area with.
Allocations are done on the basis of population and not area. This means that a CD is allocated to a postcode if the majority of the population in that CD is contained within that postcode. The POA classification also excludes what Australia Post call "secondary postcodes". These are post office box postcodes and some postcodes which are delivery routes which are also covered by other postcodes (a situation which often occurs in rural areas).

Postal Areas do cross State/Territory borders. Where this is the case standard Population Census products will provide data for the entire area.

Postcode of Usual Residence (POCUCP)

Postcode of Usual Residence (POCUCP) is derived from the postcode of usual residence as reported by the respondent. POCUCP can include all Australia Post postcodes (except secondary postcodes) valid at the time of the Population Census, however, if an Australia Post postcode is not reported by any respondents then it will not appear in the data. POCUCP also includes several codes used to classify uncodeable responses.

Studies done by the ABS have demonstrated that respondents do not reliably report the postcode of their residence. It appears that often respondents do not actually know the postcode of their residence or report the postcode of their post office box address or other mail delivery address. As such the data obtained from respondents is not considered to be of high quality.

POCUCP should not be confused with POA. POA boundaries are not able to be used with Population Census data tabulated using POCUCP.


DIGITAL BOUNDARIES

Postal Area Boundaries

Postal Areas are approximated by aggregating CDs that lie wholly or partly within the Australia Post postcode area on a 'best fit' basis. This method is comparable with the method used to generate CD Derived Postcodes used for the 1991 Census.

Postal Area boundaries are available at two levels of detail:
        • 1996 Census Reduced Boundaries; and
        • 1996 Census Shape Boundaries.
The dissemination of the 1996 Census Digital Boundaries is through a network of ABS appointed secondary distributors. Each secondary distributor has a non-exclusive license to distribute, and provide after-sales support for a set of nominated file formats. To obtain a list of registered ABS secondary providers ABS clients should contact ABS Client Services.

Australia Post Postcode Boundaries

These boundaries are only available from Australia Post or its authorised resellers.


MATCHING CENSUS STATISTICS TO BOUNDARIES

Because POA and POCUCP are different classifications, containing different ranges of postcodes, clients are advised to use the following combinations of classifications and boundaries:
        • POA statistics should be mapped using POA boundaries; and
        • POCUCP statistics should be mapped using Australia Post postcode boundaries.


COMPARING OTHER POSTCODE BASED STATISTICS TO CENSUS STATISTICS

Users of Census data often have trouble reconciling their own postcode based statistics with Census statistics. Their are a number of considerations when conducting this type of analysis, some of these considerations change depend on whether the user is comparing their data with Census POA or POCUCP data.

The overwhelming difficulty that is faced by users of postcode based data, the Census included, is that postcodes are designed for delivering mail and not for geographically referencing statistics. A number of problems are encountered when postcodes are put to this use. Firstly, it is very difficult and expensive to effectively and accurately code respondents to their "correct" postcode. Secondly, it is also difficult to say what the "correct" postcode for some respondents should be, especially when considering post office boxes and rural delivery services. Census uses postcode of residence, however, many of our users only record the postcode of a persons mailing address in their databases. The number of people with different mailing and residence addresses will always causes a lot of disparity between data sets. Thirdly, postcode boundaries do change over time and it often takes along time for respondents and coding systems to adjust to a change in a postcode. This causes problems for users working with postcode based data that is taken from different time periods as the changes to the boundaries are difficult to reconcile with changes in the population.

Comparing with POA

The primary consideration when doing a comparison with Census data based on the POA classification is that the areas are designed using the "best fit" of CDs, therefore, where the CD boundaries do not match the Australia Post postcode boundaries a degree of inaccuracy is introduced. The POA classification also excludes some postcodes and as such it will not be possible to totally reconcile these excluded postcodes if they are included in the non-Census database.

Comparing with POCUCP

When comparing POCUCP data the primary consideration is that the data is based on postcode as reported by the respondents and is therefore open to the inaccuracies associated with this type of reporting.