Table of Contents
Of the three components of demographic change, geographic mobility is the most nebulous and difficult to measure, and yet it is the one with potentially the greatest impact on population distribution and composition. Difficulties of measurement arise because a variety of definitions of population movement can be construed, all of which constitute arbitrary functions of the distance and length of time involved in relocating from one place to another. Impacts on distribution arise because migrant numbers in and out of a given place could exceed births and deaths, especially in small geographic areas and at higher stages of demographic transition, while age and sex selectivity of migration places a wide ranging effect on the composition of migrant sending-and-receiving regions.
In illuminating these issues, analysis of population movement for both the Indigenous and total Australian population has generally been informed by data yielded from fixed-period usual residence questions in the national census (that is, usual residence now compared to one and five years ago), as well as by information from case studies focused on particular localities or sub-populations (Bell 1992; Bell & Maher 1995; Gale 1972; Gale & Wundersitz 1982; Kinfu 2005; Taylor & Bell 1996a). Admittedly, between 1970 and 1987, the ABS also conducted an annual survey of internal migration in conjunction with one of its monthly population surveys (without an Indigenous identifier) (ABS 1987). However, since that time, few official nationwide surveys have included a mobility indicator in Australia. Even those that have included a mobility indicator do not have an Indigenous identifier or have limited Indigenous samples (such as the Australian Housing Survey and the February supplement of the Labour Force Survey on labour mobility).
To this extent, the inclusion of a question on mobility status and reasons for movement in the 2002 NATSISS was a rare event and it presents two new opportunities for analysts. Firstly, it provides a new source of data with which to validate previous findings from census analysis and case studies. Secondly, the existence of wide ranging individual and household characteristics data allows for testing of the statistical relationship between these variables and mobility, thereby enabling a wider and more direct exploration of the social and economic determinants of movement than hitherto possible. The purpose of this chapter is, therefore, to make use of the survey data to examine the intensity and correlates of spatial mobility among Indigenous Australians. The distinguishing feature is that the determinants of mobility are explored from a micro rather a macro perspective, while reasons provided by respondents for their movement are also explored—something that has only previously been established indirectly by proximate cause using census data (Kinfu 2005).
Interestingly, the background to the inclusion of mobility questions in the 2002 NATSISS appears to be exactly the opposite to the experience of the 1994 NATSIS. In consultations before the 1994 survey, the topic of ‘location and mobility’ was proposed as important for inclusion by a number of government agencies under the broad heading of ‘culture’. In particular, it was felt necessary to acquire data on the movement patterns of household members over a 12-month period before the survey, including the number of moves, duration of each move and reasons for each move. This mobility history was deemed by agencies such as ATSIC, Department of Employment, Education and Training (DEET), and the Department of Health, Housing and Community Services to be of assistance in planning the location of appropriate services such as community infrastructure and social programs. However, due to recall problems, less than adequate responses to mobility questions were obtained in the pilot survey and it was determined that an improvement in data quality would require overly-indulgent in-depth probing on an already crowded interview schedule. This led to the mobility questions being omitted, with the exception of a question on the number of moves away from the local area for the treatment of health problems (Taylor 1996: 41).
By contrast, in the preparations for the 2002 survey round, questions on mobility were initially excluded in line with 1994 practice. However, they were then included following strong user demand for inclusion, as mobility was viewed as a cross-cutting issue that impacted on many areas of concern (including health, education, employment and housing). While this view of the policy significance of mobility is widely held and articulated, what are less clear are the precise policy questions to be addressed by an understanding and measurement of movement. In short, what is it about mobility that we want to know and can a sample survey, such as the NATSISS, provide the answer?
The decision to include mobility questions turned out to be easier than deciding how to formulate appropriate questions. At the census, and in the now defunct ABS internal migration survey, people are classified as having moved if the address of their usual place of residence was different from that of one year or five years ago. In the 2002 NATSISS, the definition of movement is far more inclusive and refers to all moves made over the 12 months before the survey. In particular, the NATSISS asked (in non-remote and remote non-community sample areas), ‘In the last 12 months, have you lived (emphasis added) in any other dwellings?’ and (in remote community sample areas), ‘In the last year have you lived (emphasis added) in any other houses or places?’. In both areas, the survey then went on to elicit how many dwellings, houses/places people had lived in over the course of the year.
Why these differences in the nature of questions deployed on community sample and non-community sample forms arose is not known. However, leaving this aside, the key point to note is that the probability of movement should be considerably higher in the 2002 NATSISS compared to the census, given its reference to all moves as opposed to a single specific move as defined by the fixed-period census ‘usual residence’ question. This lack of any temporal or spatial reference in framing the NATSISS definition of mobility is unusual, since duration and distance of movement are two key variables for mobility analysis (Bell 2001; Taylor & Bell 2005). As mentioned earlier, the NATSISS also collected information on the main reason for the last move and classified these under four major headings: family, housing, employment and accessibility (to services). The issue of whether these questions provide an adequate and meaningful basis for addressing key policy issues—either independently or in association with other survey variables—will be dealt with in conclusion.