Two Tales of a City: detecting socio-economic disadvantage in an ‘advantaged’ Australian urban centre

Abstract

ABSTRACTUnderstanding the distribution of socio-economic status (SES) within and between communities of varying spatial scales is vital for equitable and effective resource delivery and policy development. In Australia, the Socio-economic Indexes for Areas (SEIFA) have been used extensively for these purposes, in both research and government service delivery. As area-based averages, however, their use in these areas is potentially confounded by the ecological fallacy and the modifiable areal unit problem (MAUP), statistical issues which have widespread implications for the use of spatial data. The Socio-economic Index for Individuals (SEIFI) is an individual-based index developed by the ABS to quantify SES at an individual level for 15–64 year-old Australians. Previous work investigated the potential disparities between these area- and individual-level indices, showing that considerable disparities exist, particularly in the Australian Capital Territory (ACT). Using spatial autocorrelation and pair-wise proportion analyses, we show that there was considerable diversity in individual-level SES at a very fine level of geography in 2006, a unique pattern in Australia. We also show that the SES ranking of areas within the ACT is, in many instances, not indicative of the proportions of individuals living within that area experiencing high levels of relative socio-economic disadvantage. We show that between 65–95 per cent of the ACT's disadvantaged population are masked within areas ranked as advantaged, depending on the index and level of spatial aggregation. Our results suggest that the use of SEIFA as a proxy for individual level disadvantage is highly problematic at both a theoretical and practical level, and necessitates further development of an individual- or household-level index as part of the regular suite of socio-economic indices produced by the ABS.