This final section of the report provides some interpretation for, and considers the implications of, the empirical results detailed earlier. It also explores:
- existing gaps in knowledge of violent offending perpetrated by Indigenous persons;
- the limitations of the data currently available on this issue; and
- ways in which at least some of these data gaps could be addressed.
Indigenous violent offending: A summary
While this present report has not been able to canvas all of the academic studies and government inquiries dealing with Indigenous violence, it has summarised the quantitative and (to a lesser extent) the qualitative evidence currently available. In doing so, it has confirmed that, according to both police apprehension data and self-report surveys, the rate of violent offending by Indigenous persons is consistently higher than that of non-Indigenous persons, with Indigenous males being strongly overrepresented in these figures. Levels of recidivism among violent Indigenous offenders (measured by re-contact with the criminal justice system) were also disproportionately high while, conversely, the time taken to recidivate was disproportionately low.
As outlined in the second section, Indigenous violence seems to be linked, either directly or indirectly, to a broad range of factors such as gender, age, Aboriginal/Torres Strait Islander identity, alcohol consumption, childhood experiences of abuse, exposure to pornography, indicators of socioeconomic disadvantage (such as education, employment and housing), and mental and physical health. Illicit drug use may also be relevant, although at present, Indigenous offenders are less likely than their non-Indigenous counterparts to report drug use or to attribute their offending to either dependency on, or intoxication from drugs.
However, disentangling the contribution of each of these factors to Indigenous violence is difficult, not only because of a lack of empirical data but also because of the complex inter-relationships that inevitably exist between them. Two broad explanations have been invoked, either explicitly or implicitly, in the literature.
One explanation views violence as a symptom of underlying problems, the origins of which can be traced back to the act of colonisation. In the words of Memmott and National Crime Prevention (2001: 11), '[t]he incidence of violence in Indigenous communities and among Indigenous people cannot be separated from the history of European and Indigenous relations'. According to this argument, the dispossession and removal of Indigenous people from their land and the attendant breakdown of traditional culture is directly responsible for the multiple disadvantages that now characterise many Indigenous communities, including poverty, overcrowding, dilapidated housing stock, high unemployment levels and high levels of family breakdown. These factors impact on the mental, physical, social and economic wellbeing of Indigenous individuals and, in turn, may contribute to the development of dysfunctional behaviours, such as alcohol abuse and violence. By implication then, the most effective way to reduce violence is to redress the multitude of socioeconomic disadvantages which Indigenous individuals and communities face.
While this line of reasoning has general validity, its relevance for current policy and strategy development has been challenged by Pearson (2001a, 2001b). While acknowledging that substance abuse 'originally got a foothold in our community because many people were bruised by history' (2001b: 4), Pearson argues that it is more important to focus on what is occurring now. In his view, alcohol abuse is currently responsible for 'a great proportion of Indigenous violence' as well as exacerbating existing social and economic disadvantages (Pearson 2001b: 20). Hence, while government policies and programs aimed at improving the living conditions of Indigenous Australians may be useful in preventing initial entry into alcohol misuse, they will not have any effect on an individual who is already addicted. 'Such individuals cannot be convinced to quit by offering a materially and socially better life including land rights, infrastructure, work, education, loving care, voluntary rehabilitation and so on. The addict will use all of these material and human resources to facilitate an abusive lifestyle' (Pearson 2001b: 5).
Instead, violence will only be reduced by confronting alcohol consumption and addiction directly. In part, this requires strategies designed to change community attitudes to alcohol misuse. These 'must be aimed at creating an environment which makes it more uncomfortable for substance abusers to continue with the abuse than to quit. There must be no more unconditional support if people don't change, there must be a material cost. And very importantly, there must be an immediate rejection of abusive behaviour by the environment' (Pearson 2001a: np).
In addition, there must be enforced or mandatory treatment for those who abuse alcohol; and Indigenous reliance on 'passive welfarism' must be redressed. In Pearson's (2001b) view, passive welfarism is the 'main determinant' of Indigenous substance abuse (and by extension, alcohol abuse) via its creation of idle time and lack of purpose and its provision of an unconditional money supply with which alcohol can be purchased.
Pearson's argument finds strong support from the empirical studies of Weatherburn, Snowball and Hunter (2008, 2006). While these aimed to predict the likelihood of being arrested rather than the likelihood of actually offending, and while they did not focus specifically on violence, the results clearly point to the important role now played by alcohol and illicit drug use in Indigenous arrests, with these two factors constituting the second and third strongest predictors of Indigenous arrests (after gender), even when the influence of other variables are controlled for. Equally important, they found no significant interaction effects between alcohol and drug abuse and factors such as welfare, unemployment or financial stress. These findings, they argue, indicate that drug and alcohol abuse, rather than being a symptom of cultural, social and economic disadvantage, constitute problems in their own right and play an independent role in explaining Indigenous contact with the criminal justice system and in perpetuating Indigenous socioeconomic inequality.
However, the Weatherburn, Snowball and Hunter (2008, 2006) studies also provide some support for ecological models of causation or what (using Pearson's terminology) could be designated as the 'symptom theory' of Indigenous violence, as evidenced by the fact that being unemployed rather than employed, experiencing financial stress and being dependent on welfare were all strongly and independently predictive of Indigenous arrests. These findings suggest that tackling Indigenous socioeconomic disadvantage, while not necessarily reducing violence in the short term, may nevertheless have long term benefits in this area.
Several results from another empirical study also have implications for future policy or strategic development. In attempting to predict the likelihood of violent re-offending among Indigenous violent offenders in Western Australia, Allan and Dawson (2002) found that those factors that potentially predict one form of violence differed in some respects from those that predict another form of violence (Allan & Dawson 2002). This provides a useful reminder that violence is not homogenous, but encompasses a range of different behaviours that occur in widely varying situations, target different victims and potentially have quite different triggers. To date, however, little attempt has been made to examine how the risk factors for Indigenous persons vary from one type of violence to another and how, in turn, intervention strategies and programs need to be designed to reflect these differences. The other interesting finding from this study was that one of the key predictors of the likelihood that an Indigenous violent or sexual offender would re-offend was the absence of feasible release plans at the time of exiting prison. Although tenuous, this highlights the need for criminal justice and related agencies to provide effective and timely interventions for violent offenders. There is now ample qualitative evidence to indicate that responses to Indigenous violence by mainstream agencies are, in many instances, inadequate. Failure to locate police, welfare and other critical personnel within the communities themselves reduces their ability to protect the victims and to provide effective intervention and treatment programs for offenders.
Finally, there is some evidence (albeit limited) that violence itself is a risk factor for other forms of disadvantage. According to a study by Hunter and Borland (1999), for example, Indigenous persons who have contact with the criminal justice system are significantly less likely to obtain employment than those who have never had contact. This points to the existence of a vicious, mutually reinforcing circle, whereby socioeconomic and other forms of disadvantage lead to violence, which in turn perpetuates the socioeconomic disadvantage. The challenge for policymakers is to identify where and how best to intervene in order to break that circle.
Key gaps in our understanding of Indigenous violent offending
As noted many times in this report, despite the considerable number of research and government inquiries into Indigenous violence, there is still relatively little quantitative data on the actual level and nature of violence perpetrated by Indigenous offenders or on the personal and situational characteristics of these individuals compared with Indigenous non-offenders. Instead, most of the available statistics relate to Indigenous contact with the criminal justice system rather than offending per se. Without a more detailed understanding of what proportion of the Indigenous population actually commits acts of violence, the nature and frequency of that violence, and the circumstances within which it occurs, successful intervention strategies will be difficult to develop.
Such information can best be obtained from large-scale surveys of randomly selected individuals conducted across a range of Indigenous communities. At the present time, the most suitable vehicle for collecting such information—the NATSISS—only asks individuals whether they have been arrested or imprisoned, not whether they have offended. Nor does it collect information on violence per se, or its various subcategories. In the absence of such self-report data, it is necessary to rely on official crime statistics and in particular, on police apprehension data. Apart from the obvious limitations of such data (eg they exclude the potentially large numbers of perpetrators who are never 'caught' by police and provide no comparative information on those Indigenous persons who never offend), public access to such information is comparatively restricted.
- Only two states (Western Australia and South Australia) publish regular statistics on the number of Indigenous persons apprehended by police for violent offences. It is therefore not possible to develop a national profile of Indigenous violent apprehensions or to determine how and to what extent these vary from one jurisdiction to another.
- Only one state (Western Australia) has published regionally-specific breakdowns on Indigenous violent apprehensions but even these are not released on a regular basis, with data for the 2001 calendar year being the most recent set available. The lack of regional, subregional and community-specific data prevents detailed analysis of the extent to which patterns of Indigenous violent offending vary according to the different historical trajectories and current socioeconomic characteristics of particular communities.
- No state publishes data specific to Indigenous perpetrators of family violence or child abuse and neglect on a regular basis. Nor do official police apprehensions data contain any information on those individuals who perpetrate emotional, financial or psychological abuse.
- No data are available on the extent to which perpetrators of one form of violence also commit other types of violence and/or non-violent crimes, such as property or drug offending. While research indicates that, within the general community, perpetrators of family violence may also commit child abuse, there are no published statistics specific to Indigenous offenders that can verify this.
- Data on the offending trajectories and criminal careers of Indigenous violent offenders are also lacking. Hence, determining what interventions are required and when those interventions should be applied in order to disrupt those trajectories cannot be determined.
- Very little is documented about the victims of Indigenous perpetrators of violence, including the extent to which such offenders target non-Indigenous victims or their relationship to those victims. Extrapolating from Indigenous victim reports about the offender is not the same thing, particularly in light of some data suggesting that approximately half of all assaults committed by Indigenous persons are directed against a non-Indigenous victim.
While the above gaps in our understanding of the nature and extent of Indigenous violence are important, potentially more critical is the absence of strong, empirically-based evidence of the way in which the various risk factors relate to Indigenous violence and how these vary depending on the type of violence involved and the setting within which that violence occurs. While the work of Hunter (2001), Weatherburn, Snowball and Hunter (2008, 2006) and Allan and Dawson (2002) has started to address this gap, it is still limited by their need to focus on contact with the system, rather than on actual offending behaviour, and on the limited range of potential risk factors which they are able to include. Again, because of a lack of empirical information, it is also not possible for these studies to identify how predictive factors vary depending on the different types of violence involved or on the characteristics of specific Indigenous communities.
As Weatherburn, Snowball and Hunter (2008) highlight, there are always shortcomings in research instruments, and they refer to the different time horizons in the NATSISS questions that formed the basis of the independent and dependent variables and to the wording of the question used to generate the drug abuse variable. Importantly, they call for further research into the various factors that their research showed were linked to the risk and frequency of research, in order to better understand how and why these factors have an effect. Hunter (2001) stresses that empirical analysis of large-scale survey data will continue to be only 'broadly indicative' and calls for case studies or ethnographic approaches to illuminate cross-cultural issues, including contact between police and Indigenous Australians.
How can these knowledge gaps be addressed?
There are at least three ways in which some of the current gaps in knowledge of Indigenous violent offending can be addressed:
- redesigning some of the questions currently included in the NATSISS and similar population surveys;
- making more effective use of existing data, particularly police apprehension data; and
- improving the quality of police apprehension data.
Reassessing the questions currently included in the NATSISS and similar surveys
Population-based surveys specifically targeted at Indigenous respondents provide the best method for obtaining data on the actual levels (and types) of violent offending perpetrated by Indigenous persons because they have the potential to identify all offending incidents in which the individual is involved, even if they are never identified or apprehended by police. Equally important, they also have the capacity to collect a wide range of data on the proximal factors (such as drug or alcohol use, mental impairment and stress levels) as well as on the distal factors (such as family and community characteristics) for violence. Such data would provide the basis for identifying those variables that are independent predictors of violence when controlling for the effects of other factors. Moreover, if survey numbers were sufficiently large, they could also provide the basis for detailed regional and subregional comparisons of both the levels of violent behaviour and how these levels vary depending on the characteristics of the community itself. However, a range of ethical concerns related to, for example, self-reporting of offending and the naming of specific communities, will need to be considered and addressed to ensure there is community support for such surveys.
Other data sources lack the capacity to collect such information. For example, self-report surveys that target specific groups of offenders, although useful in other contexts, only provide information on those individuals who actually offend (or have contact with the criminal justice system). While they permit comparisons between Indigenous and non-Indigenous offenders, they tell us nothing about differences between Indigenous offenders and Indigenous non-offenders, which is the key to identifying key risk factors for violence.
Apart from the problems identified earlier, police, via apprehension data, are restricted in their capacity to collect detailed information on the offender, their family environment and community setting. There have been limited attempts in some states to extend the amount of background data collected on offenders. For example, SA police have added an indicator to their Police Information Management System which, when fully operational, should allow officers to record whether, in their judgement, alcohol, illicit drug use, mental impairment or gambling are associated with that individual's offending behaviour. However, given the already heavy data recording demands placed on operational police and other justice personnel, it is unlikely that these systems will be modified sufficiently to enable them to collect all the data required to test for risk factors.
Making more effective use of existing data, notably police apprehension data
The implementation of more comprehensive and broader community-based surveys is likely to be extremely resource intensive and will require considerable lead-in time, even assuming agreement by key stakeholders that such modifications were required. A more immediate and achievable strategy for addressing some of the knowledge gaps is to make greater use of that data collected by criminal justice agencies, particularly police apprehension information.
A range of offender-based information currently collected by police is not publicly released. However, police departments have the capacity to undertake specific data extracts which, if made available to bona fide researchers, could answer some of the questions raised earlier. For example:
- In addition to Western Australia, some other states (such as South Australia) are able to extract offender-based information according to either the residential address of the offender and the location where the offence occurred. Such data would add considerably to the limited amount of regionally-specific information currently available and provide the basis for comparing rates of violent apprehensions both only within and across different types of Indigenous communities.
- Some states have the capacity to link police apprehension data with victim-based incident report data, thereby enabling them to identify and profile those offenders who, although officially charged with a generic offence such as assault, are actually targeting family members, elderly persons or children.
- States such as South Australia and Western Australia are able to identify all apprehensions involving the same individual over relatively long time periods, thereby providing some insight into issues such as whether violent offenders are charged with other types of crime and/or multiple types of violence. Data on recidivism levels could also be extracted, as well as the individual's age at the time of first apprehension, the types of offences initially committed and whether, over time, these became more serious. Techniques for determining differences in criminal trajectories are now widely used and could easily be applied to Indigenous violent offenders.
In outlining these possibilities for additional data extracts, two points should be noted:
- Only a small number of states actually have access to the data required, thus precluding any nationally-based analysis.
- Such data extracts still relate, not to all Indigenous offenders, but only to those who actually come into contact with the criminal justice system.
However, despite these limitations, such data would still be useful.
Improving police apprehension data
To improve police apprehension data on Indigenous offending it is important for all states to implement a standard process for ascertaining and recording the offender's and the victim's Indigenous status, using the ABS standard Indigenous identification question (ABS 1999). This is under active development and all jurisdictions have now agreed in principle to follow this approach. Due to the resource implications involved, it may take time for all jurisdictions to implement the changes. Once implemented, all jurisdictions will have the capacity to extract and publish Indigenous-specific information on both victims and offenders, as well as enable comparisons across jurisdictions.
Should there be more emphasis on protective factors?
The overwhelming majority of reports into Indigenous violence rely on research that focused on describing the apparent risk factors for violence and/or contact with the criminal justice system (eg Hunter 2001; Memmott & National Crime Prevention 2001; Weatherburn, Snowball and Hunter 2008, 2006). But is this the most fruitful line of inquiry? An alternative, complementary and potentially more constructive approach is to concentrate on identifying those factors that prevent or 'protect' the individual from becoming involved in violence in the first place, or once involved, will help him/her desist from such behaviour.
Despite evidence of a disproportionately high level of Indigenous violence, the fact remains that the majority of Indigenous people are not violent, even though many live in communities where violence is endemic, are subjected to violence either as victims or witnesses, and experience intergenerational trauma and systemic social disadvantage without becoming offenders themselves. In a similar vein, even within the same region, one Indigenous community may have low crime rates while another may have high rates (Lawrence 2007). What is it about particular individuals or particular communities that make them more resilient?
In the Pathways to Prevention report (National Crime Prevention 1999a) it is argued that an individual's development is marked by a number of pathways. At key transition points along those pathways, such as the transition from home to school, from pre-school through primary to high school, and from school to work, the individual may move either towards or away from offending depending on what risk and protective factors are impacting on him/her at the time. These risk and protective factors will vary from one point along the developmental pathway to another and from one time to another, depending on the changing circumstances of the individual, his or her family and the broader community. This approach therefore rejects the idea of a static list of risk and protective factors in favour of dynamic interactionism:
It is the cumulative total and the timing of adverse factors, their interactions (over time) with each other and with positive features in the environment, and the life phases involved that are more important than the compilation of lists (Homel, Lincoln & Herd 1999: 184).
By focusing on the individual's life trajectory and by identifying the critical transition points in that trajectory, interventions may be developed that divert the individual away from offending. But this approach requires a focus on both protective as well as risk factors—which is known as 'developmental prevention'.
One of the few studies to consider potential protective factors for Indigenous offending (Homel, Lincoln & Herd 1999) suggested the following:
- cultural resilience—the notion of cultural resilience arises from the 'cultural, economic, locational and structural heterogeneity' that characterises Indigenous communities across Australia and which 'represents important cultural resilience, revival and distinctiveness' (Altman 1996: 11). This idea is supported, at least in part, by findings from the NATSISS (ABS 2002). This survey indicated that, despite a long period of disruption caused by European colonisation and the upheaval of the stolen generations, indicators of Indigenous cultural retention had remained stable since 1994. Just over one-half of Indigenous people surveyed continued to identify with a clan, tribal or language group, while 22 percent continued to live in homelands and traditional country. Almost seven in 10 Indigenous people aged 15 years and over had attended cultural events in the previous 12 months while one in five still spoke an Aboriginal or Torres Strait Islander language. Successful land title claims may also have a protective role (Homel, Lincoln & Herd 1999). These claims, it is argued, 'strengthen Aboriginal communities by giving them a voice, coalescing individuals and groups, and giving rise to strong Aboriginal organisations which in turn provide sites of resistance and stronger definitions of community' (Edmunds cited in Homel, Lincoln & Herd 1999: 190);
- personal coping and adjustment skills—the survival of Indigenous communities in the face of the negative impact of colonisation and the contemporary effects of disadvantage testifies to the resilience and adaptability of individual members of these communities. Some writers have observed that Aboriginal child rearing practices tend to produce children who are self-sufficient and resourceful (Malin et al. cited in Homel, Lincoln & Herd 1999). Other research has indicated relatively high levels of self-esteem and confidence among young Indigenous males (Lincoln et al. cited in Homel, Lincoln & Herd 1999) which could provide the basis for individual resilience and personal achievement; and
- family linkages—strong extended kinship ties across generations provide a measure of economic and psychological security as well as community cohesion even if, at times, they may also be a source of friction. Such ties operate not only in more traditionally oriented, remote communities but in highly urbanised settings as evidenced by 'patterns of mobility within extended kin networks' (Daly & Smith cited in Homel, Lincoln & Herd 1999: 191). One example of this comes from a study of Indigenous families conducted in Adelaide in the 1980s (Gale & Wundersitz 1982). It found that, through a series of residential moves made over a period of years, relatively large numbers of siblings and their families managed to relocate within a few streets of each other. Given the absence, at that time, of access to vehicles and telephones, such co-location provided ready access to a strong support network in times of stress. Another coping skill identified by that study was the tendency for multiple family units from the same extended kin network to reside in the one household. The combined income of all members ensured that, as a unit, the household remained above the poverty line. The arrangement also generated an ample supply of adults who could share responsibility for supervising the young children.
If greater attention was paid to identifying the protective factors associated with violence, it may be possible to develop effective crime prevention and reduction strategies focused on strengthening these positive elements. If implemented in collaboration with Indigenous communities themselves, these may provide a more dynamic and constructive way forward than simply attempting to redress or mitigate the risk factors.
These potential protective factors can inform strategies and initiatives that draw on developmental and community crime prevention principles. However, as Hunter (2001) notes, there is rarely an explicit crime prevention objective to the multitude of early intervention projects and services that operate across the country. More careful analysis is required of social policies and programs to assess whether they are like to reduce violent offending by individuals and within communities by being sensitive to the constellation of risk factors associated with age, sex and geographic location. Given the high mobility of Indigenous Australians (see Hunter 2001), an important issue is to make policy initiatives and service delivery responsive to individuals and families that may move around to several residential locations within a year and which build on identified protective factors.
Several studies have highlighted pragmatic measures that might reduce violence in the short-term. In particular, Hunter (2001) and the later studies by Weatherburn, Snowball and Hunter (2008, 2006) drew attention to how alcohol abuse is strongly and independently associated with Indigenous Australians' contact with the criminal justice system. In their most recent study, Weatherburn, Snowball and Hunter (2008) cite evidence that indicates restrictions on alcohol supply and price increases reduces crime and antisocial behaviour in Indigenous communities. In addition to tackling socioeconomic disadvantage through national funding and policy frameworks, at a more local level, the research findings point to the potential of environmental design (housing, lighting, amenities etc) and specific employment schemes (such as CDEP) to reduce levels of violence or disorder in neighbourhood or community settings.