Observation of the 2006 IES in the large Aboriginal town of Wadeye and its hinterland uncovered a range of structural issues concerning ABS interactions with community representatives, the strategies adopted to engage, train and supervise an Indigenous census workforce, the logistics deployed to connect with a mobile and scattered population and the categorisations and interpretations associated with the census questions, which are likely to have impacted on the successful outcome of census goals. At one level, this is surprising, since many of the contingencies faced were entirely predictable. At another level, it is not, because it reflects a continuing lack of meaningful engagement between citizens to be counted and the nation-state.

Despite the cultural importance in the Aboriginal world of the area between the Daly and Fitzmaurice Rivers, from a non-Aboriginal perspective this was one of the least-known parts of the continent until the mid 1930s and numbers resident there were simply guesstimated for prewar censuses and then incorporated into the general estimate for the full-blood Aboriginal population for the entire Daly River Census District. All this began to change with the establishment of the Catholic mission in 1935, first at Wentek Nganayi, then at Port Keats (now Wadeye) in 1939. From the very outset, a key task of mission administration in the region was regular census-taking. This was a requirement in the postwar years as part of annual reporting, initially to the Native Affairs Branch, then, from 1953, to the Welfare Branch of the Northern Territory Administration (Taylor 2005). Subsequently, the official count of the population has been sourced via the five-yearly ABS census. From 1976 to 1996, this provided a count of individuals present at Wadeye on census night, with those at outstations simply included as part of a much larger number representing the balance of the entire Daly Statistical Local Area. For the 2001 Census, however, outstations located in the Thamarrurr region were identified collectively for the first time as an Indigenous location. As we have seen, in the Realpolitik of community funding and representation, the TRC has also of late engaged in enumerating its population. Since census-taking is clearly not new to the region, why then—from observation—does it seem so difficult to accomplish?

One of the ABS responses to suggestions for improving the IES was to focus on form redesign. While the effect of changes made in this regard could be established at Wadeye only a priori, there was no doubt that the new census form proved to be a highly practical instrument to administer—leaving aside issues to do with the actual questions on the form and their interpretation. In all likelihood, it would not be wrong to suggest that the single-form approach was of greater assistance to interviewers and respondents than the multi-form approach in working through what was an increasingly crowded census schedule. This, as it turned out, was the least of the issues at stake (for further comments, see Appendix B).

The fact is, no matter how good the form structure, in small-scale communal settings such as those found typically in remote Australia, other more structural and systemic issues dominate. These relate to levels of community preparedness, participation and sense of ownership in what is a substantial and highly visible interaction with government. If these are not the primary focus and concern of the IES then the task of enumeration is reduced to a direct encounter between the ABS and individual householders—a much more difficult task than working in full partnership with their representative organisations. This is not least because the IES methodology insists on a dwelling-count approach to capturing a mobile population, which heightens the need for a sufficient, knowledgeable, authoritative and experienced local census workforce. The way to secure such expertise is to strengthen the existing relations between the ABS and communities and ensure the participation of representative organisations in statistical matters as a continuing priority throughout the entire inter-censual period.

The thwarted attempts to assemble an adequate census workforce at Wadeye provide a case in point. These attempts were not assisted by a growing feeling of detachment from the census process on the part of community representatives, by disruptive communal tensions leading up to the census, by competing local demands for skilled workers and by the fact that Wadeye was just one port of call on the CFO’s vast administrative canvas. Before dismissing these as constraints that were unique to Wadeye, my point is that they were entirely to be expected and likely to be increasingly systemic in the absence of drastically improved resourcing for remote community services and governance. The consequence was an effective census team of just four people to cover the largest Aboriginal settlement cluster in the Northern Territory. This resulted in a much longer than planned for enumeration period, difficulty in establishing census coverage and a loss of capacity for data-quality checking in the field.

In the inter-cultural world of remote Aboriginal communities, the idea that a single CFO plus an Assistant can successfully negotiate, instigate and manage the census enumeration across a vast area in a compressed period without full local support is fanciful. Rather than suggesting a case for more CFOs, however, the more radical solution here is to build greater capacity for continuing ABS relations with community organisations so that by the time the census comes around every five years, both parties are better positioned to work together in the process. Part of the problem at Wadeye, it would seem, was that the census came to be viewed solely as an ABS activity—much more so than in the mainstream census, where the responsibility is on the individual householder to self-enumerate, and increasingly so given online census access. In the IES, errors can arise precisely because the householder has far less control over the process and is much more dependent on the logistical capabilities of ABS officers and procedures (this is not to deny that the IES, as conceived, is a device to enhance Indigenous participation in the census). To foster closer collaboration, or partnership, it might help to pursue notions of rights and responsibilities around census-taking in much the same way that Shared Responsibility Agreements (SRAs) work for other areas of government activity. For example, at Wadeye, ABS access to the TRC population database for guidance on DOB data and data-quality checking might have been more likely to have eventuated if ground rules and modus operandi regarding confidential data access were negotiated more fully in advance.

Ultimately, the census has two broad objectives. The first is to measure accurately the number and key characteristics of people in Australia on census night and the dwellings in which they live. The second is to provide timely, high-quality and relevant data for small geographic areas and small population groups, to complement the rich but broad-level data provided by ABS surveys (ABS 2006a). The IES, of necessity, contravenes the simultaneity condition of the first of these objectives, and by so doing compromises its capacity to deliver on the second. The dominant finding from observation of events at Wadeye is that this structural weakness of the IES could be greatly ameliorated if Indigenous people and their representative organisations were adequately positioned and resourced to engage more meaningfully in the census process and thereby assume more—not less—responsibility for ‘self-enumeration’ in line with the rest of the Australian community.