Experiences of racism among Aboriginal and Torres Strait Islander adults living in the Australian state of Victoria: a cross-sectional population-based study

Background

The history of Indigenous people in Australia over the last two hundred years since the arrival of Europeans is one of great suffering. Genocide, the introduction of European diseases, dispossession, subjugation and segregation reduced the Indigenous population by 90% between 1788 and 1900 [1]. A conservative estimate indicates that prior to European contact there were approximately 15,000 Indigenous Australians living in the state of Victoria; that number was reduced to approximately 850 by 1901 [2].

Today, Indigenous Australians continue to face interpersonal and institutional racism which creates and sustains their lower socioeconomic status by excluding them from economic opportunities and land ownership. Moreover, Indigenous men and women can expect to live 10.6 and 9.5 years less than non-Indigenous men and women respectively [3]. A large and growing body of evidence consistently implicates racism as a key determinant of the health of Indigenous Australians [4, 5].

However, one of the most persistent aspects of today’s discourse regarding racism in Australia is the very denial of its existence [6]. A review of the linguistic and discursive patterns of contemporary speech in both informal and formal (parliamentary debates, political speeches, and the media) settings in Australia concluded that the social taboo against openly expressing racist beliefs has led to the development of strategies that present negative views of minority groups as reasonable and justified, while exonerating the speaker from charges of racism. This serves to constrain political efforts to address racism thus reinforcing racism [6].

In this paper we sought to: (a) quantify the population-based prevalence of racism experienced by Indigenous Australians who lived in the state of Victoria; (b) determine if experiences of racism among Indigenous Victorians are independent of lifestyle risk factors and social determinants, such as socioeconomic status, that are often used to justify negative stereotypes; (c) describe potential pathways by which experiences of racism impacts on Indigenous Victorians to create inequalities in health; and (d) identify key points of intervention and potential strategies to combat racism.

On an individual level, racism refers to the beliefs and attitudes that members of certain groups have of their superiority in relation to other groups who are regarded as inferior, based on race, ethnicity or cultural background [7]. Those who are assumed to be inferior are treated differently and unfavourably.

At a societal level, racism can be defined as “… that which maintains or exacerbates inequality of opportunity among ethnoracial groups” and racial discrimination as the racist behaviours and practices that result in inequality of opportunity among ethnoracial groups [8].

Methods

Results

Seventeen percent of Indigenous adults experienced at least one episode of racism in the year preceding the survey, compared with 4.5% of their non-Indigenous counterparts (Table 1). Thus, Indigenous adults living in Victoria were four times more likely than non-Indigenous adults to experience racism (odds ratio (OR) = 4.3; 95% CI = 3.2–5.8).

However, Victoria is a multicultural state with people from all over the world, including a large non-white non-Anglo-Celtic population who began to immigrate to Australia after the repeal of the White Australia Policy in 1973. By excluding participants who were not born in Australia to Australian-born parents and spoke a language other than English at home, we excluded the majority of the non-white non-Anglo-Celtic population who may similarly have experienced racism. While we assumed that there would be misclassification error, in the absence of any other data on ethnicity, the majority of those born in Australia to non-Indigenous Australian-born parents who only spoke English at home are likely at this point in time to be of the dominant white Anglo-Celtic population. This reduced the prevalence of experiences of racism in the non-Indigenous population from 4.5 to 2.8% (Table 1) and the OR of the association between experiences of racism and Indigenous status increased from 4.3 to 7.2 (5.3–9.7). Thus, Indigenous adults were 7 times more likely to experience racism than non-Indigenous adults who were born in Australia to Australian-born parents who only spoke English at home.

Socio-demographic variables that were significantly associated with experiences of racism included age, sex, rurality, and three indicators of socioeconomic status (Table 1). Age was inversely associated with experiences of racism; as age increased, experiences of racism decreased. Males were more likely than females to experience racism, as were those who lived in metropolitan compared with rural Victoria.

The association of experiences of racism with socioeconomic status varied according to the measure employed. Living in a household with a total annual income of less than $40,000, not being tertiary-educated, and not being in the labour force (unable to work, retired, engaged in home duties, or student) were associated with a lower prevalence of experiences of racism. In contrast, adults who were unemployed were almost twice as likely to experience racism as those who were employed.

Social support and trust are measures of ‘social capital’. There is no single definition of social capital. However, in essence social capital refers to the nature and extent of one’s social relationships across society, which determines access, or lack thereof, to the social and economic resources needed for a good life.

When we investigated the relationship of experiences of racism with social capital, we found that adults who were unable to get help when needed, irrespective of the source of help, were significantly more likely to experience racism than adults who were able to get help from any of these sources (Table 3). Similarly, social and civic trust were also associated with experiences of racism. Adults who did not believe that most people could be trusted were almost 4 times as likely as those that did believe most people could be trusted to experience racism; OR = 3.6 (3.1–4.1). Similarly, adults who did not feel safe walking alone down their street after dark, those who did not feel valued by society, and those who did not feel there were opportunities to have a real say on important matters, were more likely to experience racism.

When we investigated the relationship of experiences of racism and lifestyle risk factors, we found that adults who smoked, were underweight or obese, and who were physically inactive, were significantly more likely to experience racism than non-smokers, people of normal weight, and the physically active (Table 2). We used two measures of alcohol consumption and found a u-shaped relationship: Indigenous adults who abstained from alcohol consumption and those that drank excessively on any given occasion were both more likely to experience racism.

When we controlled for each secondary independent variable in a bivariable analysis; age, education, and social trust were the only three variables that changed the OR of the association between experiences of racism and Indigenous status by more than 10% and were thus deemed to be potentially explanatory of the association (Table 4).

In our study, twice as many non-Indigenous adults (44%) were aged 65 years and older compared with their Indigenous counterparts (22%). Controlling for age reduced the OR by 23% from 4.3 to 3.3 (2.5–4.5).

Similarly, a higher proportion of Indigenous adults (22%) than non-Indigenous adults (13%) did not believe that most people could be trusted. Controlling for social trust, reduced the OR by 12% from 4.3 to 3.8 (2.8–5.1).

In contrast, low educational attainment appeared to be protective against experiences of racism, as controlling for education increased the OR by 11% from 4.3 to 4.8 (3.6–6.5).

When we included all secondary independent variables in a multivariable analysis, the OR was reduced by 29% from 4.3 to 3.0 (2.2–4.3). However, the association between experiences of racism and Indigenous status remained highly significant at the p < 0.001 level (Table 4).

Controlling for the three variables deemed to be potentially explanatory (age, education, and social trust) in a multivariable model, reduced the OR by 21% from 4.3 to 3.4 (2.5–4.7).

Controlling for all social determinants reduced the OR by 26% from 4.3 to 3.2 (2.3–4.5). In contrast, controlling for all lifestyle risk factors only reduced the OR by 10% from 4.3 to 3.9 (2.9–5.2).

Discussion

The prevalence of experiences of racism among Indigenous adults who lived in Victoria between 2011 and 2014, was 17% (13.3–21.5%), compared with 4.5% of non-Indigenous adults. Indigenous adults were four times more likely to experience racism than their non-Indigenous counterparts (OR = 4.3; 3.2–5.8). However, compared with the largely white non-Indigenous population of Anglo-Celtic origin, Indigenous adults were seven times more likely to experience racism (OR = 7.2; 5.3–9.7) as only 2.8% of Anglo-Celtic adults reported that they had experienced discrimination or been treated unfairly because of their racial, ethnic, cultural, or religious background. We expect that the estimate of 2.8% is likely to be an overestimate because some of these experiences of discrimination may have been due to religious background rather than race, ethnicity or culture.

Although we know that non-Indigenous adults of non-Anglo-Celtic origin also experience racism, our interest was specifically in the Indigenous experience of racism because of the enormous health inequities that exist between Indigenous and non-Indigenous adults. Whereas, the non-Indigenous non-Anglo-Celtic tend to be recent migrants who have better health than those born in Australia; commonly referred to as ‘the healthy immigrant effect’ [15].

To the best of our knowledge, we believe this study to be the first population-based study of experiences of racism among Indigenous adults who live in the state of Victoria.

However, we believe that our estimate of experiences of racism among Indigenous Victorians is a significant underestimate of the true prevalence of racism and that the estimate of 17% should be taken as meaning ‘at least’ 17%. We say this for the following reasons. Firstly, Indigenous status was determined by a simple single-item that asked, “Are you of Aboriginal or Torres Islander origin”. Some people may say yes to this question because they have a distant relative who was/is Indigenous, but they personally do not identify as Indigenous and may not ‘look’ Indigenous, given that there remains a widespread erroneous belief that indigeneity is about skin colour. Therefore, these individuals may not be at risk of experiencing racism due to their self-reported Indigenous origin. We have no way of distinguishing or quantifying such participants. If they made up a significant proportion of the Indigenous sample, then the prevalence of experiences of racism would be significantly underestimated. Indigeneity in Australia is generally determined by a three-part definition that must be met to be legally considered to be Indigenous. A person must have Indigenous heritage, identify as Indigenous and be accepted as such by an Indigenous community [16].

Secondly, the wider literature consistently shows that experiences of racism are typically under-reported [17]. Studies show that people are more likely to report experiences of racism if the question is phrased to ask about the experiences of the ethnoracial group to which they belong, rather than their personal experiences [17–19]. Evidence suggests this may be due to the psychologically protective effect associated with minimising personal experiences of racism [20, 21]. The Victorian Population Health Survey only inquired about a participant’s personal experiences of racism.

Thirdly, multi-item measures of experiences of racism tend to be more reliable than single-item measures [22]. For example, the 2014–15 National Aboriginal and Torres Strait Islander Social Survey, a population-based national survey that used a multi-item measure of experiences of racism, reported that 34% of Indigenous Australians experienced racism [23]. By contrast, the Victorian Population Health Survey only used a single-item measure.

It is important to be cognisant of the fact that racism is a complex phenomenon and reducing it to a single-item question cannot capture its complexity [24]. Indeed, its prevalence is highly likely to be underestimated when using a single-item question. Moreover, Indigenous people view racism as a more diverse and complex phenomenon than non-Indigenous people [25].

It is also well-known that survey questions developed for one culture may not be culturally appropriate for another culture. In recognition of this and the complexity of the phenomenon of racism, Paradies and Cunningham (2008) developed, tested and validated a 31-item Measure of Indigenous Racism Experiences (MIRE) to assess experiences of racism among Indigenous Australians [25]. Future research on the prevalence of experiences of racism among Indigenous Australians will likely be more accurate if it utilised the MIRE questions.

In 2011, the Localities Embracing and Accepting Diversity (LEAD) survey, conducted in Victoria, reported that 97% of Indigenous participants experienced racism [26]. The purpose of the LEAD study was not to specifically measure the prevalence of racism, but to investigate the relationship between experiences of racism and mental health outcomes. As a result, this study was not population representative of Victoria, as it was conducted in only four localities of Victoria and recruitment was non-random to maximise recruitment. However, as this study was conducted among specific Indigenous communities who met the three-part definition of indigeneity and almost all participants had experienced racism, its findings support our contention that our estimate is an underestimate of the true prevalence of racism experienced by Indigenous Victorians.

The time period of exposure to racism is also of importance. Our study asked about the previous 12 months while another study inquired about the lifetime prevalence of exposure to racism and estimated that 52.3% of Indigenous urban Victorians aged 12–17 years experienced racism [5].

Whether the prevalence of experiences of racism is higher or lower in Victoria compared with other states is currently unknown. However, we hypothesise that there may be a higher prevalence of experiences of racism in Victoria because Victoria has the lowest ethnic density (0.9%) of Indigenous Australians than any other state and there are only two discrete Indigenous communities in Victoria, which have small populations [27]. High own group ethnic density has consistently been shown to be protective against experiences of racism, believed to be due, at least in part, to a lower exposure to the perpetrators of racism [28–30].

Controlling for social determinants and lifestyle risk factors attenuated, but did not eliminate, the strong statistical association between experiences of racism and Indigenous status. Indigenous Victorians were still at least three times more likely to experience racism than their non-Indigenous counterparts after controlling for these factors.

The social determinants had a larger impact on the association between experiences of racism and Indigenous status than the lifestyle risk factors. The negligible impact of the lifetyle risk factors refutes the commonly made claim that the racism Indigenous Australians experience is due to their ‘poor lifestyle choices’, rather than their Indigenous status [7, 31]. This is particularly pertinent when considering alcohol consumption, as there is a commonly held negative stereotype that most Indigenous Australians drink alcohol to excess, often used to justify racism [31]. However, the evidence shows that Indigenous Australians are less likely to consume alcohol than non-Indigenous Australians and we found that Indigenous Victorians who abstained from alcohol consumption were as likely to experience racism as those who drank excessively [32].

Experiences of racism varied by socioeconomic status. Indigenous adults of low socioeconomic status, whether measured by household income, educational attainment or not being in the labour force, were less likely to experience racism than their higher socioeconomic counterparts. The single exception was that those who were unemployed were also more likely to experience racism. Our findings are consistent with the literature. For example, the Darwin Region Urban Indigenous Diabetes (DRUID) study also found a higher prevalence of experiences of racism among Indigenous Australians of high socioeconomic status [33].

A possible explanation of why experiences of racism are higher among Indigenous Australians of higher socioeconomic status may be that those who manage to overcome the substantial barriers that Indigenous people continue to face in mainstream society are a minority within a minority. This is likely to increase an individual’s exposure to mainstream society and put them at a greater likelihood of experiencing racism, which is consistent with the evidence on the protective effects of high own group ethnic density. It may also explain the seeming contradiction of unemployed Indigenous Victorians being more likely to experience racism. Unemployed Indigenous Victorians may also have a higher exposure to mainstream society because such exposure is necessary to receive unemployment benefits. Alternatively, or additionally, it is possible that people of higher socioeconomic status have a greater propensity to report experiences of racism.

Our findings that low educational attainment appear to be protective against experiences of racism is of concern given the poorer socioeconomic outcomes associated with low levels of education. There is a large body of research demonstrating the existence of maladaptive problem-focused behavioural responses to racism, such as opting out of formal education as an act of self-protection [34]. This may help to explain the lower secondary school completion rates among Indigenous children and is supported by a recent study in Victoria, which identified racism within the school system as one of the most challenging issues faced by Indigenous children, particularly at secondary school level [35].

The implication of this finding is that more needs to be done to eradicate systemic and interpersonal racism within our education system. In 2012, the Race Discrimination Commissioner observed that: “Sometimes racism can be reflected in not telling the history of an event or the experience of a group of people in our country” [36]. Currently, what is, or is not, taught in schools about Indigenous history and culture, depends on individual schools. Unfortunately, an attempt to introduce a national curriculum, which embedded education about Indigenous culture, history, and the impact of colonisation, was thwarted in 2014 by the Federal government, following a non-independent review of its content [37].

At a societal level, groups who claim ethnoracial superiority at the expense of those deemed inferior, derive great benefits from the inequitable social and economic living conditions that are generated [17]. However, for the group deemed to be inferior, chronic experiences of racism are harmful to their mental and physical health [38–40]. While racism is not always intentional and much of systemic racism is carried out by people who are ignorant or in denial, this doesn’t lessen its harmful effects.

According to Krieger’s ecosocial analysis, the harm occurs through seven pathways [17]: (1) economic and social deprivation; (2) higher exposure to toxins, hazards, and pathogens; (3) social trauma, (4) health-harming responses to racism, (5) targeted marketing of harmful products; (6) inferior and inadequate health care; and (7) environmental degradation and alienation from the land [22].

Strengths of our study include that it was based on data from the Victorian Population Health Survey, which has been conducted annually since 2001 and is a well-validated population-based survey with a relatively high response rate. Moreover, the Victorian Population Health Survey collects data on a wide breadth of topics, including the social determinants of health, because it was informed by a public health model of the social determinants of health [41]. In contrast, most health surveys tend to be informed by the biomedical model of health, which attributes disease to proximate biological factors at the level of the individual and ignores the social determinants of health. Collecting data on the social determinants of health provdes an opportunity to develop policy directions that address racism.

Weaknesses of the study, other than those previously described, include the use of two different questions about experiences of racism. While the questions in the 2011 and 2012 studies ask specifically about experiences of racism directly attributable to Indigenous status, the question in the 2014 survey asked about experiences of racism attributable to ‘racial, ethnic, cultural, or religious background’. This leaves open the question of potential intersectionality between race and religion and whether we are accurately measuring experiences of racism among participants from the 2014 survey which would impact on the prevalence estimate of the combined dataset. The prevalence of experiences of racism among Indigenous participants from the 2011 and 2012 surveys was 19.6% (13.8–27.1%) compared with 15.0% (10.7–20.7%) from the 2014 survey. Although lower among the 2014 Indigenous participants, the difference was not statistically significant.

Survey data is cross-sectional, which does not allow for conclcusions to be drawn in relation to cause and effect or its direction. For example, feeling unsafe walking alone after dark could be the consequence and/or cause of self-reported racism.

The data is self-reported raising concerns about bias and accuracy. However, not all data readily lends itself to objective measurement, and experiences of racism are an example [34]. However, it is self-reported racism that is strongly associated with mental and physical ill-health [42].

The Victorian Population Health Surveys conducted prior to 2015 only surveyed households with landline telephone connections. Yet the exponential uptake of mobile telephones has caused a rapid growth in households that rely solely on mobile telephones and raised concerns that telephone surveys that only include fixed landline connections are losing their population representativeness [43]. Moreover, Indigenous women have been found to be five times more likely than non-Indigenous women to live in mobile-only households [43]. Therefore, if the experiences of Indigenous households who have landline telephones are different to those who do not, our findings may not be as population representative as we suppose.

As noted by the extensive work of Maggie Walter, the collection, analysis and interpretation of data on Indigenous peoples are not as objective as non-Indigenous peoples claim them to be [44]. In Australia most research is conceived, conducted and interpreted by non-Indigenous people who are largely of middle class Ango-Celtic origin. Consequently, the research decisions made reflect the social norms, values and beliefs of the non-Indigenous. This has lead to a lot of research that effectively stigmatises Indigenous people, thus reinforcing racism.

For example, there is a disproportionate amount of research that focuses on health behaviours such as smoking and alcohol consumption, comparing Indigenous with non-Indigenous people. Such research concludes that Indigenous people are more likely to engage in unhealthy health behaviours than their non-Indigenous counterparts, which is stigmatising [45]. This has lead to policies aimed at closing the gap in health between Indigenous and non-Indigenous Australians being almost exclusively focused on reducing the gap in health behaviours, which is notoriously hard to do in any population. Yet the irony of this is that health behaviours only account for approximately 32% of the total burden of disease and this itself may be an overestimate as it is based on a study that only included health behaviours in the risk factors analysis [46]. It is the social determinants of health that account for a far greater proportion of the burden of disease.

The reasons for this not only reflect the dominance of the biomedical model of health, which was conceived in Europe and the United States, but also Western neoliberal culture that values individualism over collectivism and regards individual responsibility as the pathway to good health. It is at odds with the Indigenous perspective on health. Imposing such beliefs and values through prioritising this type of research is, arguably, racist. We therefore recognise this as a weakness of our study and join the growing call for better engagement with, and the inclusion of, Indigenous people and Indigenous researchers at all stages of the research process, from conception to publication.

Each year the Prime Minster of Australia reports on progress towards closing the gap in Indigenous health. However, in the 9 years since the commencement of the ‘Closing the Gap’ strategy, very little has been achieved and in some cases the gap is widening [47]. The National Indigenous and Torres Strait Islander Health Plan 2013–2023, designed to address the gap, acknowledges that “racism is a key social determinant of health for Indigenous and Torres Strait Islander people…” and seeks to eliminate systemic racism within the healthcare sector. However, it still disproportionately focuses on changing the health behaviours of Indigenous Australians and ignores the wider systemic and interpersonal racism directed against Indigenous Australians [48]. We contend that the gap is unlikely to be reduced until we comprehensively address racism towards Indigenous Australians [49].

A large body of work on anti-racism strategies and interventions has been conducted and trialled by the Victorian Health Promotion Foundation [50]. We refer readers to their website [50]. Table 5 attempts to summarise potential policies and interventions, by sector, that may effect real societal change in attitudes and behaviours. The list is not meant to be exhaustive, but rather to provoke thought. Many of the policies and interventions are aimed at eliminating systemic racism rather than interpersonal racism which is the subject of this study. However, all forms of racism ought to be tackled simultaneously to prevent reversion. Moreover, piecemeal approaches to tackling racism that are often underfunded and not sustained have the potential to do more harm than good [51].

In Australia, a paternalistic ideology continues to pervade policy-making for Indigenous Australian across all levels of government [52]. This is, therefore a key area for reform. Paternalistic policies are inherently racist as they do not recognise the right to self-determination and seek to limit the choices of individuals, based on the belief that individuals do not know what is in their best interest. The antithesis of the paternalistic approach is a human rights-based approach. The adoption of a human rights-based approach to policy-making would be more likely to facilitate the elimination of systemic racism which in turn would lead to better health and wellbeing outcomes for Indigenous people.

Conclusions

This study shows that, contrary to the current discourse in Australia that denies the existence of racism, racism directed against Indigenous adults in Victoria is a significant problem and may be associated with lower educational attainment, which may lead to lower socioeconomic status and poorer health outcomes.

Therefore, if as a society we truly wish to reduce the gap in health between Indigenous and non-Indigenous Australians, we should: (a) acknowledge that racism against our Indigenous counterparts exists; (b) that it is extensive and harmful; and (c) that it is a major determinant of the gap in health. Moreover, racism directed against Indigenous Australians is a problem that needs to be addressed by the dominant non-Indigenous population through challenging and changing beliefs and behaviours in schools, workplaces, the media, the public sector, government and society at large.

Authors’ contributions

AM analysed the data and drafted the article. ZA, DC, and JM provided information, contributed to the draft, read and approved the final version of this article. All authors read and approved the final manuscript.