The 2002 NATSISS provides the first survey data on Indigenous residential mobility indicating the propensity to move with reasons for doing so. Brief analysis of these data confirms many of the findings from previous census-based analysis: that the probability of movement peaks among young adults, is similar overall for males and females, is higher for single people and especially high among the unemployed, greater for those in private rental dwellings, and lowest in remote areas. Among the new findings are factors that tend to deflate mobility. These include low educational attainment, home ownership, residence in community housing, residence on their homeland, living in an area with neighbourhood problems, and outer regional location. Also new is the insight that CDEP employment does not dampen movement propensity as previously suggested (Taylor & Bell 1996b). While this much can be gleaned, and while further outcomes and relationships will no doubt be established, even the preliminary findings presented here raise a number of issues regarding the utility and interpretation of NATSISS data on mobility.

First of all is the surprising outcome that the overall propensity to move is only marginally higher than reported in the census, and that in line with the census, movement rates are lowest in the Northern Territory and remote areas generally. So pervasive is the observed fact of residential movement in the daily, fortnightly, seasonal and annual round of Indigenous social and economic life in remote Australia, as elsewhere in the country, that ethnographers have been strained to describe it using evocative terms. These include beats, runs, lines, floaters, visitors, concertina households, or multi-locale relationships (Taylor & Bell 2004). If the 2002 NATSISS fails to reflect the manifest intensity of population movement in remote areas, then questions naturally focus on non-sampling error and, in particular, how the notion ‘lived in’ might have been presented by interviewers and interpreted by respondents. Having an open-ended question as in the NATSISS is potentially valuable for capturing short-term and repeat movement, but without spatial, temporal, and conceptual structure it can descend into ambiguity.

This brings us to the purpose of a mobility question in a survey such as the NATSISS. In pressing the case, many agencies in submissions to the ABS highlighted the likely impact of mobility on their portfolio areas of concern. If there was ever an expectation that the NATSISS might measure the spatial impact of population movement for the purpose of planning service requirements (the most useful insight from a policy perspective) then this was ill-informed. One assumes, however, that more abstract higher-order questions were in mind, such as those answerable by multi-variate analysis. But again, when we are dealing with movement defined in such a way that it that could be for one night next door as opposed to 11 months across the continent, or from Kintore to Alice Springs instead of Newcastle to Sydney, then structurally—and policy-wise—we are dealing with very different types of mobility that remain totally undifferentiated in the NATSISS. Having said that, further work could be done to isolate sub-groups in the survey population that present potential interest for policy (such as those who indicated employment reasons for movement, or particular difficulties with housing), and their characteristics and behavioural attributes could be explored in more detail. However, much of this is speculative after the event and, as with all such surveys, clear specification of the answers to be addressed by end-users is essential for optimal design.

If we turn to the actual reasons for movement provided by respondents, these conflict (to some degree) with the findings of the multi-variate analysis which highlights structural (economic) factors such as unemployment and private rental housing as key predictors of movement. It is also the case that accessibility to services scores very low as a reason for movement, which is contrary to the experience of many Indigenous people who have to move in order to access basic services such as health care, schooling, training, banking or shopping. The fact is, separating out reasons for mobility is never easy given the multi-faceted purpose and unpredictable nature of many journeys—for example, what may have started out as a shopping expedition to town can turn out to be an extended stay with relatives (Young & Doohan 1989). Nonetheless, it is interesting that Indigenous people choose to emphasise social over economic factors (such as employment), while many housing reasons (such as overcrowding and wanting a bigger house) can be viewed as both social and economic. This emphasis on social factors is consistent with much of the ethnographic research on mobility that underscores the role of kinship networks in directing population flows, even if the primary purpose might be job search or housing (Gale 1972; Gale & Wundersitz 1982; Gray 2004).