Explaining the Census: investigating reasons for non-response to the ABS Census of population and housing

This paper introduces some work currently in progress at the Institute for Social Research at Swinburne University of Technology. The study forms part of an Australian Research Council Linkage Project being undertaken by Swinburne in partnership with the Australian Bureau of Statistics (ABS). The objective of the Linkage project is to improve the design and management of census operations in Australia, using methods drawn from the disciplines of Statistics and Information Technology. It will do this by developing a statistical model that can explain patterns of response and non-response in the 2006 census, and predict patterns of response that appear in 2011. Of particular interest to the project are localities where the level of 'dwelling non-response' â- a concept defined in Section 2 of this paper â- is high compared to others in their state. The research undertaken in the Linkage project is the first of its kind in Australia. While some information about 'undercount' â- another concept defined in Section 2 â- is provided in ABS publications, there has been no detailed analysis of the ways in which levels of non-response to the census are related to other factors such as demographic characteristics, socio-economic measures or location. Investigation of these questions is happening progressively as the project proceeds. No findings from this work have as yet been published. After early analysis identified unexpectedly high levels of dwelling non-response in inner-city areas in Melbourne and in other states, the Institute for Social Research was contracted to do a qualitative study to find out why this might be so. This publication is a background paper for that study.