Table of Contents
The National Census is a broad-brush instrument with two major objectives: to provide an accurate count of the national population and to collect data on demographic and socioeconomic characteristics that are comparable across different sectors of the population, variously defined. The Indigenous Enumeration Strategy (IES) has evolved through the years in response to the perceived ‘difference’ of the Aboriginal and Torres Strait Islander populations of Australia. In the 2006 Census, we were in a unique position to observe the workings of the IES, from the design of the collection instrument through to the processing of the data collected in the field. Our findings suggest that the IES has probably reached a point in its development where the injection of ever-increasing resources into essentially the same generic set and structure of activities could begin to produce diminishing returns to output (data quality) unless there is some fundamental reworking of the way in which the strategy is delivered. In particular, we suggest that the manner in which the Australian Bureau of Statistics (ABS) engages with local communities and their organisations needs to change.
We begin with a short discussion of the Interviewer Household Form (IHF) before turning to what emerged as the major theme of our 2006 research: the difficulties of using a dwelling-based count that assumes a sedentary lifestyle to capture a highly mobile population—a population that manifests little consciousness of or interest in the purposes of the census. We will frame this discussion in terms of agency—the agency of Indigenous people and the organisations that sit on the interface between local Indigenous populations and the state—and contingency—the need to accept the contingent factors that influence Indigenous mobility and to build a response to contingency more explicitly into the collection strategy. We argue that to contain the effects of contingency it is necessary for the ABS to make much more productive use of local knowledge at the regional and the local level. In part, this entails building the agency of local institutions into the process much more effectively.
As indicated in Chapter 1, evaluation of the IHF was initially a primary objective of the research since the IHF—as a single matrix form—was considerably different from the two-form collection instrument of 2001. In all four case-study accounts (Chapters 3–6), there is a sense that the new IHF (see Appendix A) worked quite well. The most obvious of the remaining structural problems is the ordering of the questions on persons temporarily absent (PTA) before the questions about who was at the dwelling—residents and visitors—at the time of the count. We discuss this issue—which could have had a substantial effect on the adequacy of the count—in Appendix B. All four case studies also contain comments about questions that did not work very well for various reasons. These too are discussed in detail in Appendix B.
After the census count, the observation of the final checking of census forms at the Darwin Census Management Unit (CMU) and the conversion of paper-based returns to computer-read unit record data at the Data Processing Centre (DPC) in Melbourne allowed us to follow raw data all the way from collection to processing. This has provided a unique basis for evaluating census output. Questions such as how household structures are compiled, how occupations are determined and what contributes to final head counts have never before been analysed in this way. While we note the intense, often highly technical, scrutiny of census returns that uses a mix of formulaic criteria and almost forensic guile, at the end of the day there remains an inevitable sense of human agency in decision-making about certain data categorisations, not least because the overarching framework for this in terms of the ultimate purpose of data collection remains, inevitably, a function of state administration.
Observation of the family-coding procedures (Chapter 8) confirmed our findings in 2001 (Morphy 2002, 2004, 2006; see also Morphy 2007) that the census cannot capture fully the complex family structures of Indigenous households and in many ways it is better if it does not try to do so. We conclude that the 2006 questions about relationships within households (Questions 13, 30 and 31) worked better than the 2001 questions, but they could still be improved somewhat (see Appendix B). We argue again also for the recognition of a new coding category of ‘extended-family household’.
In the case of occupation coding, the current checklist for the majority of employed Indigenous people in remote Australia working on a Community Development Employment Projects (CDEP) does not adequately identify all economic activity. A significant proportion of the remote-area Indigenous economy therefore remains invisible in the 2006 Census.
These observations on coding point to the very partial way in which the census captures Indigenous social and economic life in remote Australia. We return to this question in a postscript to this chapter.