Very wisely, the DPC had instituted a dedicated ‘pick list’ for Community Development Employment Project (CDEP) occupations.^[7] The majority of ‘employed’ Indigenous people in remote Australia are CDEP participants (see, for example, Gray and Chapman 2006: 117). The instructions to the coders were to use this pick list as the first port of call whenever CDEP was mentioned, no matter which of the employment-related questions it appeared in. They were also warned that some people who worked for CDEP organisations were not CDEP participants, and to exit the CDEP pick list into the mainstream occupation coding options when a person’s job description evidently did not fit within the CDEP listings. There was a concern that sometimes the level of CDEP-related jobs might be understated. For example, if someone said that they were a ‘receptionist’, this should not be automatically coded to ‘administration or clerical work’ on the CDEP pick list.

The CDEP pick list was compiled using responses from the 2001 Census, and this might have been an adequate strategy were it not for the case that CDEP was undergoing rapid and radical change at the time of the 2006 Census (DEWR 2005a, 2005b, 2006). The process of refocusing CDEP towards training people for ‘real jobs’ was encouraging the relabelling of ‘traditional’ CDEP jobs. Women who had previously been receiving their CDEP for ‘home management’ or ‘home duties’ were now working in ‘environmental services’ or ‘community care’, for example, as were many former ‘rubbish collectors’. These changes to CDEP were very recent at the time of the census, and in Arnhem Land I had observed some people still using the older categories. By the time of the next census, if CDEP is still in existence, I would expect most CDEP participants to be more aware of their ‘official’ job descriptions, since each CDEP participant is now given a written formal job description and training plan. It would be advisable then to construct the CDEP pick list primarily from the Department of Employment and Workplace Relations’ (DEWR) list of job descriptions in 2011, possibly supplemented by information from the 2006 responses.

The increasingly significant role of the Indigenous Protected Area (IPA) scheme and associated ranger programs was evident in the responses to CDEP questions. The closest equivalent in the CDEP pick list was ‘park ranger’. It was decided to best fit ‘community ranger’—the most common description of ranger jobs associated with IPAs—to ‘park ranger’ rather than simply picking it from the list, as a way of alerting the data-analysts to the gap in the pick list. By 2011, depending on the continuing success of the IPA program and the ability of these ranger programs to attract funding, many of these ranger positions could be non-CDEP jobs.

The instruction to use CDEP as the entry to job classification wherever it appeared in the answers to work-related questions was a relaxation of normal coding procedures—but a sensible one. In general, respondents are not able to discriminate very clearly between questions that ask for a distinction between a job description and the list of tasks performed, nor are they able, in many cases, to provide a coherent or accurate description of what their ‘employer’ does. The questions on the census form are designed with mainstream employment situations in mind, and do not have salience for people who are not employed in the mainstream. Had the coders been faced with attempting to code CDEP responses as if they were mainstream responses, their task would have been much more arduous and frustrating.

As in 2001, I was struck forcibly by the silences—the gaps in information—that the form of the census questions on economic participation engendered. The overwhelming impression of remote Indigenous Australia is of economically marginalised people, in low-paid, part-time, low-skilled, ‘dead-end’ occupations. The biggest silences concern the subsistence economy and participation in the art industry. Some commentators (Hughes 2007; Johns 2006) dismiss subsistence hunting and gathering as merely recreational, whereas in many areas of remote Australia—particularly in outstation communities—there is evidence that these activities contribute significantly to the health and wellbeing of the population (see, for example, ABC News Online 2006; Barber 2005; McDermott et al. 1998; Morice 1976; Watson 2007). There is, however, nowhere on a census form where people are encouraged to record these activities—not even in the question on unpaid domestic work. Perhaps for the next census that question could be reworded to make it less ‘feminine’ and ‘hunting, fishing or gathering bush food’ could be included among the suggested options.

The Indigenous arts industry is a major component of the remote economy (see, for example, Altman 2003). The study area where I observed the enumeration in 2006 contained one extremely successful art centre, many successful and well-known artists and many others who were learning from the more established artists. Despite the inclusion of ‘artist’ in the list of suggestions for occupations, however, only one or two people put this down as their occupation. It is clear that the majority of people do not really think of art production as a job, and the emphasis in the census questions on sources of regular income—as opposed to intermittent and somewhat unpredictable income—conspires with this attitude to render an important source of income and economic engagement essentially invisible.

From the Yolngu point of view therefore, art production is not a ‘job’, even if it is a source of income. There are, however, some things considered by Yolngu to be ‘jobs’ that the mainstream categories ignore. Several senior Yolngu who were CDEP participants described their occupation as ‘leader’ or ‘community leader’. Needless to say, there was no match for these in the CDEP pick list, and it is very unlikely that such are their ‘official’ CDEP job descriptions. It proved very difficult to deal with such cases even in ‘best fit’ or by exiting the CDEP module, since equivalent jobs in the mainstream generally imply high levels of formal (Western) education and qualifications, and incomes to match. From the point of view of these senior Yolngu, however, this is indeed their most important job, and it can be demanding and time-consuming.

It is extremely interesting to see these cultural differences in the idea of what is and is not ‘work’ being played out through the census process. What is frustrating is that none of this complexity is evident in the official output of the census—the tables representing the ‘facts and figures’ of Indigenous employment. Arguably, however, if effective measures are to be put in place to improve the economic circumstances of Indigenous people, it is as necessary to understand differences in perceptions and values as it is to understand the ‘facts’ as they appear through the lens of the mainstream.