LITERATURE REVIEW FINDINGS IN BRIEF
Note that the word ‘Indigenous’ is used throughout this discussion of the literature as it pertains to internationally comparable conceptions of Indigenous populations.
A range of research work, both academic and governmental, was assessed to ascertain the current understanding of differential identification both in Australia and overseas. Other countries with large Indigenous populations (USA, Canada and New Zealand) were included, though barriers to direct comparison are noted. Broadly, the literature review found that propensity to identify is discussed as a factor influencing Indigenous statistics in all four countries considered in the review. To varying extents, differential identification patterns (specifically, ‘new identifications’, where individuals who have not previously identified as Indigenous do so for the first time) are thought to have contributed to non-demographic growth in Indigenous populations in recent decades.
Changing propensity to identify impacts on population counts and on the assessment, for the purposes of social policy, of the needs of the relevant population. There are two key ways in which this is observed. If propensity to identify has a greater impact on either Census counts or administrative data (that is, if the effect of propensity to identify is not uniform across data sets and across time), changes in the population count could create a change in rate statistics. Rate statistics use the population count as the denominator to calculate the frequency of an occurrence (for example, the proportion of Aboriginal and Torres Strait Islander people who have completed Year 12) relative to the whole population. If the denominator (the population count) changes because of something that does not impose the same change on the numerator (the variable of interest, e.g., Year 12 completion), the rate statistic changes. This ‘denominator shift’ could create the impression of a change in some outcome variables when no such shift has in fact occurred in the population being measured (Barnes, 1997). Independent of this, individuals identifying for the first time may share certain characteristics. If a large number of people with a particular characteristic ‘appear’ in the population for the first time, this could create a change in the measure of that characteristic in the total population. That is, the aggregate measure of that characteristic may change without any real change having occurred in that variable for the population that was last measured.
The exact amount of non-demographic population change that can be accounted for by changes in propensity to identify is uncertain. Variability in propensity to identify has nonetheless been recognised by the ABS and other research bodies as a key factor in population variability (ABS 2002; ABS, 1998; ABS & AIHW, 2003; Ross 1996). The ABS has attempted to incorporate differential identification into its population projections, issuing a ‘low series’ estimate based on the assumption of no change in identification (assuming population change on the basis of measurable demographic factors only) and a ‘high series’ estimate, which allows for some variability in identification patterns (ABS, 1998).
Factors that are proposed as contributing to propensity to identify include marriages between Aboriginal and/or Torres Strait Islander people and people of non-Indigenous descent, geography and the social environment in which identification occurs. People who have one Indigenous and one non-Indigenous parent are observed as having particularly variable identification behaviours both in Australia (Ross, 1996) and internationally (Chapple, 1999; Gould, 2000), while individuals living in urban areas often represent the populations in which high levels of non-demographic growth are observed (Ross, 1996). Changes in individual propensity to identify were observed in a New Zealand longitudinal study (Carter, Hayward, Blakely & Shaw, 2009); survey respondents reporting multiple ethnicities at wave 1 were among the most likely (alongside those reporting sole Maori ethnicity) to report different ethnicity in subsequent waves.
Changes in the social environment over time, in particular since the civil rights movements of the 1960s, are cited as contributing to increases in Indigenous identification (Guimond, 2006). Other changes in policy and public opinion are also linked to high rates of Indigenous population growth. This is particularly noticeable in Canada, where changes to legislative definitions of Indigenous groups and the benefits available on the basis of group membership have preceded marked growth in those groups (Guimond, Kerr & Beaujot, 2003).
While not solely focused on Indigenous populations, the work of Duncan and Trejo (2005) on social mobility and inter-marriage highlights a key issue of consideration for ethnic groups – that individuals who marry outside their ethnic group report better economic outcomes, and their children are more variably identified. New Zealand researchers have also found that people who have one Indigenous and one non-Indigenous parent appear to achieve better socioeconomic outcomes than those who solely identify with a disadvantaged sub-population (eg. Maori) (Chapple, 1999; Gould, 2000). Assuming that propensity to identify is most variable in populations with high rates of exogamy, shifts in identification in this sub-population (the population of people with mixed Aboriginal/Torres Strait Islander and non-Indigenous heritage) could have impacts on aggregate outcome measures for the overall Aboriginal and Torres Strait Islander population.