Foreword | Conferencing and cautioning are used as diversionary alternatives in the juvenile justice system and there is evidence to suggest they reduce reoffending. As Indigenous young people are overrepresented in the juvenile justice system, an important question is whether they are as likely to be diverted as non-Indigenous young people. This study used modelled data to examine juveniles' contact with the police and courts, and the differences in juvenile diversionary rates for Indigenous and non-Indigenous offenders in New South Wales, South Australia and Western Australia in 2005. For all states, Indigenous young offenders were more likely than non-Indigenous offenders to be referred to court, non-Indigenous offenders were more likely to receive a police caution, and males and older offenders were more likely to be diverted. The number of prior contacts was similar for all states, with more contacts reducing the likelihood of diversion and with less likelihood of diversion for offenders committing offences against a person. As Indigenous young offenders are more likely to have multiple prior contacts with the system, including detention, further research is needed into the reasons for their high reoffending rates.
Judy PuttGeneral Manager, Research
The overrepresentation of Indigenous young people in the criminal justice system is one of Australia's most significant social problems. This overrepresentation was highlighted by the Royal Commission into Aboriginal Deaths in Custody as a contributing factor to the rate of Indigenous deaths in custody (Commonwealth of Australia 1991). Although significant resources have been allocated to reduce this overrepresentation, the problem still remains. In 2005, Indigenous juveniles accounted for 52 percent of 10 to 17-year-olds in juvenile detention across Australia. In Western Australia (WA) the corresponding rate was 75 percent, in South Australia (SA) 44 percent and in New South Wales (NSW) 52 percent (SCRGSP 2007).
New South Wales, South Australia and Western Australia have introduced systems of conferencing and/or cautioning to reduce the overall rate of juvenile contact with the criminal justice system. In New South Wales, this is referred to as a Youth Justice Conference (YJC), in South Australia a Family Conference (FC) and in Western Australia a Juvenile Justice Team (JJT). Conferences are facilitated by a trained conference convenor. Family members of the offender, the victim(s), members of the criminal justice system and other interested parties can attend, along with the offender. The offence and its impact on the victim(s) and the wider community are discussed, and the offender is encouraged to accept responsibility and negotiate some form of restitution (also see 'Note').
Although these diversionary alternatives appear to be effective in reducing reoffending (Luke & Lind 2002), Indigenous young offenders appear to be much less likely to be diverted than their non-Indigenous counterparts. Data from South Australia indicate that Indigenous young offenders are more likely to be sent to court and less likely to receive a formal caution or be diverted to a FC than non-Indigenous young offenders (Wundersitz & Hunter 2005).
Similarly, in Western Australia, Indigenous young people are five times more likely to have had formal contact with the police and 29 times more likely to have been arrested (in the 10-14 year age group) (Loh & Ferrante 2003). In New South Wales, Indigenous young people are more likely than non-Indigenous young people to be taken to court (64% compared with 48%) and less likely to be cautioned (14% compared with 28%) by police (Chan et al. 2004).
The reason for this discrepancy in rates of juvenile diversion is of critical importance to criminal justice policy. Studies by Luke and Cunneen (1995) and Cunneen (2006) have argued that racial bias in the exercise of police discretion early in the criminal justice process contributes to Indigenous overrepresentation in juvenile detention centres and prison. Their argument is that because Indigenous young people are more likely than non-Indigenous young people to be arrested rather than cautioned, they tend to acquire a more extensive criminal record at a young age. The possession of a longer criminal record then increases their risk of detention or imprisonment when they reappear in the criminal justice system, even if they appear for offences that are comparable to those committed by non-Indigenous offenders.
This is a plausible hypothesis; however, there are other possible explanations. It is possible, for example, that Indigenous offenders are less likely to be diverted simply because they less frequently meet the legal requirements for diversion.
This paper compares Indigenous and non-Indigenous juvenile diversionary rates in the jurisdictions of Western Australia, South Australia and New South Wales, to determine how much of the discrepancy between Indigenous and non-Indigenous rates of diversion remains once legally relevant factors have been taken into account.
The current study
Comparability of juvenile justice systems
Before discussing the methods employed in the current study, it is important to point out that the legislative regimes that underpin the juvenile justice systems of each of the states, though similar, are not identical and that differing legal requirements determine the application of, and eligibility for, diversionary options. The SAYoung Offenders Act 1993 (SA-YOA), for example, states that:
A charge may only be laid - (a) if the youth requires the matter to be dealt with by the Court; or (b) if, in the opinion of the police officer, the matter cannot be adequately dealt with by the officer or a family conference because of the youth's repeated offending or some other circumstance of aggravation. (s 7(4))
The SA-YOA gives no specific guidance regarding the factors that can be considered when making the decision to divert a young person. It does state that an offender has to admit the offence and the offence has to be 'minor'; however, this is not further defined.
The WAYoung Offenders Act 1994 (WA-YOA) states that the police are to caution an offender:
...unless because of the number of previous offences with which the person has been charged or for which the person has been dealt with under this Part it would be inappropriate only to give a caution. (s 23(1))
The types of offences that would render a caution inappropriate are specified under Schedules 1 and 2 of the Act and include serious violent offences (e.g. murder, sexual assault) as well as other serious offences (e.g. driving causing death). However, the number of previous offences that would render a caution inappropriate is not specified by the Act. With regard to JJTs, the WA-YOA provides some guiding principles for the application of this option (under s 29) and endorses the use of JJTs for first offenders in particular (s 29). However, an important condition on the use of JJTs is that an offender can only be diverted to a JJT if they accept:
...responsibility for the act or omission constituting the offence, and agrees to have the matter dealt with by a juvenile justice team rather than by a court... (s 25(4))
Further, if a potential participant (which could include the police, the victim or another party) in the proceedings does not agree that a JJT is appropriate, the matter cannot be dealt with by a JJT. A responsible adult also needs to be present, although the JJT can appoint someone in place of a responsible adult if it deems fit, or dispense with this requirement if the young person is deemed independent.
The NSWYoung Offenders Act 1997 (NSW-YOA) states that a formal caution or a YJC can be prescribed for a young person if the young person has admitted the offence, consented to a caution or YJC, committed an offence for which a caution or YJC can be given, and is entitled to a caution or YJC. The decision about entitlement is made regarding the seriousness of the offence, the degree of violence, the harm caused to the victim, previous offence history and any other matter the official thinks is appropriate.
There are two things worth noting about all these provisions. The first is that the discretion available to police and courts in relation to whom they choose to divert is rather wide. None of the schemes contains a complete list of the factors that police and courts must or can take into account when deciding whether to caution a young offender. This makes it impossible to determine whether differences in rates of juvenile diversion can be explained solely in terms of legal factors. The second is that, although each state requires attention to different factors, all states place considerable emphasis on:
- the offence type
- the number of previous contacts
- the type of previous penalties
- whether the offender admits the offence.
Aims of the study
The current study aimed to determine how much of the difference in rates of diversion remains after relevant demographic and offence-related factors have been taken into account. Also, the distribution by Indigenous status of characteristics relevant to the decision to divert an offender was considered to assess whether the differences in diversion could be explained in part by an offender's characteristics.
For various definitional and procedural reasons it was not possible to include jurisdiction as a variable in our analysis because, as can be seen above, the states' legislation differ in the requirements for diversion. Each of these analyses were carried out separately for all three states. The general pattern of findings was then compared for the states.
Data and methodology
The data for South Australia were obtained from the Office of Crime Statistics and Research (OCSAR), the data for Western Australia were obtained from the University of Western Australia's Crime Research Centre (UWA-CRC) and the data for New South Wales were obtained from the NSW Bureau of Crime Statistics and Research (BOCSAR). The data were drawn from both police and court records.
In this study, diversion has been defined as either a formal police caution or a referral to a relevant conferencing authority, irrespective of whether the conference was court-referred or police-referred. Other available diversionary schemes (such as drug courts) were not considered because there was not enough commonality between these schemes for the three states. Informal cautioning was also not considered because it was not available from administrative datasets. Consequently, this study is potentially underestimating true diversion.
There is considerable debate over whether a court-referred conference fits the definition of diversion. Because of these concerns, the modelling was conducted twice, using diversion as defined above and using diversion where court-referred conferences were not included in the diversion category. The results changed in magnitude but not in direction, and therefore only the results based on models where diversion includes all conferences are presented here; however, this does not affect the conclusions drawn from the results.
In summary, a 'contact' was defined as a formal police caution, a referral to a conference or a court appearance.
The dataset used in the modelling and bivariate analyses contained a record for all juvenile offenders who had a contact with the justice system in 2004. This dataset contained all offenders under the age of 18. When an offender had more than one contact, the most recent contact was chosen. The WA sample contained 8,669 offenders, the SA sample 4,664 and the NSW sample 6,827.
The outcome variable - type of contact - was regressed against Indigenous status, age and sex, and the following control variables:
- type of principal offence in the index contact. This was categorised as: offences against the person, drug offences, property offences, public order offences, traffic offences and other offences. The principal offence was defined as that which received the most serious penalty or the one with the most serious statutory penalty (if offences received the same penalty). Other offences include: offences against judicial procedures (such as breach of a justice order), weapons and explosives offences, and property damage (including graffiti).
- number of prior contacts (defined as cautions, conferences and court appearances)
- whether the offender has previously been given a custodial sentence
- the number of years since the first contact
- the age of the offender at the first contact.
Bivariate analyses were then carried out between each of the explanatory variables and Indigenous status to determine which variables contributed the most to the discrepancy in rates of diversion.
The results reported here are an abbreviated version of those contained in Snowball (2008).
Considering unadjusted rates, in all three states Indigenous young offenders were significantly more likely than their non-Indigenous counterparts to be referred to court. In Western Australia and New South Wales, Indigenous young offenders were also more likely to be referred to a conference rather than cautioned. Non-Indigenous offenders in all three states were significantly more likely to receive a police caution. The unadjusted analysis was carried out using a different dataset from that used in other analyses in this report. Refer to Snowball (2008) for details about this dataset.
The likelihood of receiving the respective outcomes was modelled as both a dichotomous (diversion/non-diversion) and ordered (caution, conference and court) variable. Snowball (2008) contains results from both modelling procedures; only the dichotomous analysis is presented here. Note that in both the dichotomous and the ordered models, the dependent variables exerted a similar effect in direction and only differed in magnitude.
Table 1 presents the odds ratios and 95 percent confidence intervals for the variables that remained significant in the final model. In all three states, the 'Age at first contact' and 'Years since first contact' variables were not significant. Due to a small number of offenders being classified as 'Indigenous unknown', this variable was not included in the WA model. For the same reason, the 'Prior custodial sentences' variable was not included in the SA model. If the effect of an offence did not differ significantly from the 'Other' base group, the model was recalculated with the 'Other' base group incorporating that offence. For example, in the NSW model, 'Public order' offences did not exert a significantly different effect from the 'Other' group. The model was recalculated and the new base group effectively became 'Other' and 'Public order' offences. In the models for all three states, the 'Age at first contact' and the 'Number of years since first contact' variables were not significant at the five percent level and hence were not included in the final model.
|Comparison||WA model||SA model||NSW model|
|Indigenous vs Non-Indigenous||0.426(0.364-0.499)||0.629(0.505-0.784)||0.497(0.427-0.577)|
|Indigenous unknown vs Non-Indigenous||-||1.162(0.856-1.577)||0.057(0.042-0.079)|
|Male vs Female||0.702(0.598-0.824)||0.786(0.661-0.935)||0.676(0.584-0.783)|
|Offences against the person vs Other offences||0.146(0.107-0.200)||0.558(0.456-0.682)||0.658(0.552-0.783)|
|Drug offences vs Other offences||1.674(1.115-2.511)||-||2.235(1.662-3.005)|
|Property offences vs Other offences||0.677(0.507-0.905)||-||2.073(1.780-2.415)|
|Public order offences vs Other offences||0.512(0.369-0.711)||2.086(1.638-2.657)||-|
|Traffic offences vs Other offences||0.165(0.122-0.233)||0.081(0.064-0.102)||0.099(0.078-0.129)|
|Number of prior contacts||0.660(0.643-0.678)||0.717(0.697-0.738)||0.655(0.627-0.685)|
Odds ratios less than 1 suggest that the variable reduces the probability of diversion, whereas those greater than 1 suggest the variable increases the probability. The further the odds ratio is from 1, the larger is the effect of the variable.
Indigenous status had a significant and strong effect in all three states. The odds ratio of less than 1 suggests that Indigenous offenders were less likely to be diverted than non-Indigenous offenders with similar characteristics that could be controlled for in the model; that is, an offender of the same sex and age who has committed an offence in the same grouping and has the same number of prior contacts and the same prior custodial sentence status.
In all three states, males and older offenders were less likely to be diverted. The number of prior contacts had a roughly similar effect in all three models and significantly reduced the likelihood of diversion. Since this is a continuous variable, the effect increased with the number of contacts. In Western Australia and New South Wales, a prior custodial sentence had a very pronounced effect on the likelihood of diversion. In both states, offenders with this characteristic were significantly less likely to be diverted.
The effect of offence type differed across the states. In Western Australia, offenders who committed offences against a person were significantly less likely to be diverted than offenders who committed 'Other' offences. Similarly, traffic offences also significantly increased the likelihood of a court appearance. Public order and property offences had a less pronounced but still negative effect on the probability of being diverted. However, drug offences increased the likelihood of diversion.
In South Australia, offences against a person and, in particular, traffic offences, had negative effects on the likelihood of diversion. 'Public order' offences significantly increased the likelihood of diversion.
In New South Wales, offenders charged with a traffic principal offence were considerably more likely to have their matters dealt with in court. Offences against a person also increased the likelihood of a court appearance. Offenders charged with property and drug offences were more likely to be diverted.
The previous section suggested that prior contacts and having had a previous custodial sentence significantly reduced the likelihood of diversion. Table 2 presents the distribution of the variables by Indigenous status. Snowball (2008) presents the distribution of other variables.
|Western Australia||South Australia||New South Wales|
|No prior contacts||36.9||66.6||36.5||53.3||29.1||58.3|
|One prior contact||12.7||14.5||11.0||15.1||17.6||17.6|
|Two prior contacts||9.6||6.3||6.7||7.7||12.1||8.4|
|Three prior contacts||6.5||3.6||5.2||4.8||9.2||5.2|
|Four prior contacts||5.2||2.1||3.8||3.3||6.9||3.2|
|Five prior contacts||4.1||1.6||4.0||2.5||6.3||2.4|
|Six prior contacts||3.0||1.4||2.1||1.7||4.1||1.5|
|Seven prior contacts||3.5||0.7||2.5||1.5||3.1||1.0|
|Eight or more prior contacts||18.6||3.1||28.2||10.1||11.7||2.5|
|Prior custodial sentence||8.0||1.1||3.1||1.2||10.8||2.7|
In all three states, there was a large discrepancy in the number of prior contacts between Indigenous and non-Indigenous offenders. In Western Australia, two-thirds of non-Indigenous offenders were making their first contact with the criminal justice system. This was true for slightly more than one-third of Indigenous offenders. Almost one in five Indigenous offenders had eight or more previous contacts. This compares with one in 32 non-Indigenous offenders. It is clear that Indigenous offenders, on average, have longer criminal histories than non-Indigenous offenders.
In South Australia, more than half of the non-Indigenous offenders were making their first contact with the justice system, but this was the case for just over one-third of Indigenous offenders. Furthermore, whereas three in 10 Indigenous offenders had eight or more contacts, only one in 10 non-Indigenous offenders had eight or more prior contacts with the criminal justice system.
The discrepancy between Indigenous and non-Indigenous offenders is particularly pronounced in New South Wales. About three in five non-Indigenous offenders had no contact with the justice system prior to the present contact. This is the case for only three in 10 Indigenous offenders. Conversely, whereas more than one in ten Indigenous offenders had eight or more previous contacts, only one in 40 non-Indigenous offenders had that number of contacts.
There was also a significant difference between Indigenous and non-Indigenous offenders for prior custodial sentences. In Western Australia, one in 13 Indigenous offenders had previously been sentenced to custody, as compared with one in 88 non-Indigenous offenders. In South Australia, the discrepancy between the rates of prior custodial sentences is not as pronounced, although Indigenous offenders are still almost three times more likely than non-Indigenous offenders to have previously received a custodial sentence. In New South Wales, more than one in 10 Indigenous offenders had this characteristic, compared with one in 37 non-Indigenous offenders.
Summary and discussion
The present study had two aims. The first was to assess the extent to which the observed differences in diversion between Indigenous and non-Indigenous offenders could be explained by relevant factors. The second was to examine differences between Indigenous and non-Indigenous juvenile offenders for characteristics known to affect the likelihood of diversion.
In all three states, Indigenous offenders were considerably more likely to be referred to a court than non-Indigenous offenders. In Western Australia and New South Wales, Indigenous young offenders were also more likely to be referred to a conference rather than cautioned. Non-Indigenous offenders in all three states were significantly more likely to receive a police caution. When controls were introduced for age, sex, characteristics of the current case and the prior criminal history of the offender, the discrepancy between Indigenous and non-Indigenous offenders in rates of diversion decreased for all three states but remained statistically significant.
It is impossible to say whether the residual differences in rates of diversion are symptomatic of racial bias on the part of police (or courts) or reflective of other factors that are unable to be measured in the present study. As pointed out in the introduction, the legislation establishing diversion schemes in each state affords police wide discretion in determining who should be referred to court and who should be cautioned or referred to a conference. The available data, for example, do not permit any assessment of whether an offender accepted responsibility for an offence, although this is plainly relevant to decisions about how to deal with a young offender. Moreover, other factors such as the lack of diversionary alternatives in regional or remote rural areas may also contribute to differential rates of diversion.
Looking at the variables that were significant predictors of diversion across all three states, it is obvious that past contact with the justice system plays just as important a role in shaping subsequent decisions about diversion as it does in shaping decisions about adult sentencing (Snowball & Weatherburn 2007). Irrespective of jurisdiction, Indigenous young offenders are much more likely than their non-Indigenous counterparts to have had multiple contacts with the criminal justice system and much more likely to have been in custody before. The modelling suggested that these two factors greatly reduced the likelihood of diversion.
The present study therefore strongly suggests the need for further research into the reason(s) for the high re-conviction rate among Indigenous offenders. The most obvious explanation for the high juvenile Indigenous reappearance rate in court is that Indigenous young people are more likely to reoffend, and there is some evidence to support this hypothesis (Weatherburn, Fitzgerald & Hua 2003). However, it is possible that other factors are in play, such as a lack of diversionary alternatives in regional or remote rural areas, a perceived lack of contrition on the part of Indigenous young offenders or a lack of resources for diversionary programs in remote rural areas where Indigenous families reside.
Regardless of the reason for the high rate of Indigenous reappearance in court, it is important to remember that diversionary policies are more likely to achieve their objective of reducing contact with the criminal justice system if they are effective in reducing reoffending. Research examining the reasons behind offending in Indigenous adults have highlighted social and economic factors such as unemployment, financial stress, lack of education, and alcohol and drug abuse as contributing factors to involvement in crime (Weatherburn, Snowball & Hunter 2006). Although no similar research has been conducted into Indigenous juvenile offending, it would be surprising if these were not also important factors.
The system of conferencing was modelled as a continuous variable and coded as 1, 2, 3, 4, 5, 6, 7, 8, 9 and 10+ for Western Australia and South Australia, and 1, 2, 3, 4, 5, 6, 7 and 8+ for New South Wales. The final category was grouped in this manner to ensure the variable remained linear against the logit of the outcome variable. For this reason, the variable can only be interpreted for 10 or fewer prior contacts for Western Australia and South Australia, and for eight or fewer prior contacts for New South Wales.
- Chan J, Bargen J, Luke G & Clancey G 2004. Regulating police discretion: an assessment of the impact of the NSW Young Offenders Act 1997. Criminal law journal 28(2): 72-92
- Commonwealth of Australia (1991). Royal Commission into Aboriginal Deaths in Custody, vols. 1 & 4. Canberra: Commonwealth of Australia
- Cunneen C 2006. Racism, discrimination and the over-representation of Indigenous people in the criminal justice system: some conceptual and explanatory issues. Current issues in criminal justice 17(3): 329-346
- Loh N & Ferrante A 2003. Aboriginal involvement in the Western Australian criminal justice system: a statistical review, 2001. Perth: Crime Research Centre.
- Luke G & Cunneen C 1995. Aboriginal over-representation and discretionary decisions in the NSW juvenile justice system. Sydney: Juvenile Justice Council of NSW
- Luke G & Lind B 2002. Reducing juvenile crime: conferencing versus court. Crime and justice bulletin no. 69.
- Snowball L 2008. Juvenile diversion and Indigenous offenders: a study examining juvenile offenders in Western Australia, South Australia and New South Wales. Canberra: Criminology Research Council.
- Snowball L & Weatherburn D 2007. Does racial bias in sentencing contribute to Indigenous over-representation in prison? Australian and New Zealand journal of criminology 40(3): 272-290.
- Steering Committee for the Review of Government Service Provision (SCRGSP) 2007. Overcoming Indigenous disadvantage: key indicators 2007. Canberra: Productivity Commission
- Weatherburn D, Fitzgerald J & Hua J 2003. Reducing Aboriginal over-representation in prison. Australian journal of public administration 62(3): 65-73
- Weatherburn D, Snowball L & Hunter B 2006. The economic and social factors underpinning Indigenous contact with the justice system: results from the 2002 NATSISS survey. Crime and justice bulletin no. 104.
- Wundersitz J & Hunter N 2005. Juvenile justice in South Australia: where are we now? Information bulletin no. 40.
About the authors
Lucy Snowball is a Senior Research Officer at the NSW Bureau of Crime Statistics and Research.
This paper is taken from the report of research commissioned by the Criminology Research Council.