Australian Institute of Criminology

Skip to content

Introduction

Indigenous overrepresentation is the most significant social justice and public policy issue within the Australian criminal justice system. Despite the existence of justice agreements and plans in every jurisdiction over the past decade, the gap has continued to widen in every jurisdiction (ABS 2012a). Indigenous people aged 10 years and over were between 5.6 and 8.4 times more likely than non-Indigenous people to be arrested during 2009–10 (ABS 2012b). Indigenous youth were 13.4 times more likely than non-Indigenous youth to be under community supervision and 23.9 times more likely to be in youth detention during 2009–10 (AIHW 2011). Indigenous adults were 14.3 times more likely than non-Indigenous adults to be incarcerated during 2011 (ABS 2012a).

Two national policy initiatives are driving attempts to reduce Indigenous disadvantage, including Indigenous overrepresentation in the criminal justice system. The Closing the Gap strategy recognises the need for a long-term approach to reduce Indigenous disadvantage (COAG 2009). The strategy aims to achieve simultaneous improvements in seven areas of life—early childhood, schooling, health, economic participation, healthy homes, safe communities, and governance and leadership. The National Indigenous Law & Justice Framework aims to create safer Indigenous communities (SCAG 2009). One of the main mechanisms proposed to reduce Indigenous overrepresentation as offenders in the criminal justice system is through the use of effective and targeted crime prevention programs. Unfortunately, little publicly available information exists regarding how programs might be targeted to reduce offending by Indigenous peoples.

Frameworks driving crime prevention

Two of the main frameworks that shape our understanding of offending and that may be used to target interventions aimed at reducing offending are the criminal careers paradigm, and crime and place. This section provides an overview of each approach, highlighting how they improve our understanding of offending and may be used to target interventions.

Criminal careers framework

The criminal careers framework has been described as one of the most visible areas of scholarship within criminology (DeLisi & Piquero 2011). Within this field, studies that focus on the nature, pattern and correlates of offending over the life course have been conducted in many jurisdictions (see DeLisi & Piquero 2011). These studies aim to improve understanding about how offending develops and factors that can potentially be manipulated to hinder initiation, hasten desistence and reduce career length (Blumstein et al. 1986; Piquero et al. 2001, 1999). Several major longitudinal studies have been carried out in the United Kingdom (Piquero, Farrington & Blumstein 2007; Jones, Nagin & Roeder 2001), United States (Chung et al. 2002; Piquero et al. 2001), Canada (LaCourse et al. 2003) and New Zealand (Fergusson, Horwood & Nagan 2000). This research has found that:

  • offending peaks in the late teenage years;
  • the peak onset age of offending is between eight and 14 years;
  • the peak desistence age of offending is between 20 and 29 years;
  • the process of desistance operates across all offenders;
  • early age of onset predicts a relatively long criminal career duration and the commission of relatively many offences;
  • there is marked continuity in offending and antisocial behaviour from childhood into adulthood;
  • a small proportion of the population commit a large proportion of all crimes; and
  • different types of offences are committed at distinctly different ages.

Criminal careers research has been aided by statistical techniques, such as the Semi-Parametric Group-based Method (SPGM; Nagin & Land 1993).The SPGM identifies different groups, each with their own trajectory, to capture the variation in offending in the data (Kreuter & Muthén 2008). In his review of over 80 studies that employed this technique, Piquero (2008) drew four main conclusions. First, research identifies at least two offender groups—an adolescent-peaked pattern and a chronic offender pattern. The chronic offender pattern includes a small proportion of offenders who account for relatively high proportions of offences. This group begins offending early in life, at high rates, and persists at relatively high rates when the norm seems to be desistence from offending. Research also typically identifies a late-onset chronic group, who begins offending during adolescence and continues offending into adulthood. Second, the trajectory method typically identifies between three and five groups, slightly more in studies using self-reports of offending than official records. Third, a sample size of greater than 500 provides robust categorisation of groups. Finally, there tends to be a low-rate group, a high-rate group and a moderate but declining group.

Knowledge derived from criminal careers research is particularly useful for understanding whether certain groups of offenders should be targeted and when interventions are likely to be most effective. While few trajectory studies have been conducted in Australia, findings indicate that there is a small group of early-onset chronic offenders who account for a large proportion of offending. This group comprises between three percent and 11 percent of offenders and accounts for 27 percent to 33 percent of offences (Allard et al. under review; Livingston et al. 2008; Stewart et al. under review). Not surprisingly, Indigenous Australians are more overrepresented in the early-onset chronic offender group than other offender trajectories. Livingston et al. (2008) found that 50.9 percent of the chronic group were Indigenous offenders, while 25.4 percent of the adolescent limited group and 18.4 percent of the adolescent onset group were Indigenous. Stewart et al. (under review) found that Indigenous peoples were 11.3 times more likely to be in the early onset chronic offender group, with 7.3 percent of all Indigenous peoples in Australia in this group compared with 0.6 percent of non-Indigenous people.

Targeting crime prevention towards potential chronic offenders is likely to be a cost-effective approach. Recent criminal careers research has assessed the costs of individuals on different offender trajectories. Cohen, Piquero and Jennings (2010a) explored costs using ‘bottom-up’ and ‘top-down’ costing approaches. The ‘bottom-up’ approach involved assessing the value of specific cost categories that result from crime, including victim costs, criminal justice system costs and the cost of forgone earnings by the offender. The ‘top-down’ approach was based on the public’s willingness to pay to reduce crime, which produces higher estimates because it includes collateral costs relating to fear of crime (ie crime prevention expenditure, avoidance behaviour and insurance costs) and loss of social cohesion. When costs were applied to individuals in the offender trajectories, the high-rate chronic offender group constituted 3.1 percent of the sample but over 40 percent of costs. Each high-rate chronic offender was found to cost either US$515,382 or US$1.1m by the time they turned 27 years of age, depending on whether intangible costs were included. In their follow-up study, Cohen, Piquero and Jennings (2010b) used a ‘top-down’ costing approach and examined costs separately based on sex and ethnicity. While a different number of trajectories were identified, 2.8 percent of males were found to be high-rate chronic offenders and they accounted for 37 percent of male offending costs, or in excess of $1.5m each. Although 0.5 percent of females were chronic offenders, they accounted for 49 percent of female offending costs or US$754,440 each. Offending by African-Americans was found to be the most expensive out of any ethnic trajectory group and averaged in excess of US$1.6m for each chronic offender.

Two studies conducted outside the United States have also assessed the costs of crime using ‘bottom-up’ costing approaches. In Australia, Allard et al. (under review) found that an early onset chronic offender group comprised three percent of offenders yet accounted for 26.5 percent of costs, with each early onset chronic offender costing $323,645 in criminal justice system and wider social and economic costs. A second chronic offender trajectory group was also identified, with adolescent onset of offending. This group comprised 1.8 percent of offenders and accounted for 15 percent of costs, with each adolescent onset chronic offender costing $302,034. Piquero, Jennings and Farrington (2011) assessed the costs of offender trajectories based on the Cambridge Study in Delinquency Development, which included convictions of 411 South London males aged 10 to 50 years. The high-rate chronic offender group was found to cost over 10 times as much as other groups, with each offender costing $US95,241.

Unfortunately, it is difficult to target potential chronic offenders because there is a lack of research that differentiates offender trajectories based on risk factors, with no Australian studies. Nevertheless, this group would be ideal candidates for developmental/early intervention. Programs based on this approach target at-risk children, aiming to reduce the number of risk factors and increase the number of protective factors (see Table 1). The effects of risk factors on development appear to be cumulative, interactive and sequential (Farrington 2002; Granic & Patterson 2006). However, the accumulation of multiple risk factors appears to be more important than the acquisition of specific risk factors for the development of offending (Farrington 2002; Howell 2003; Stouthamer-Loeber et al. 2002; Tremblay & LeMarquand 2001; Wasserman & Miller 1998; Wasserman & Seracini 2001). Evidence indicates that offending is much more likely among those who are exposed to or experience greater levels of risk, such as many Indigenous peoples (Bonta, LaPrairie & Wallace-Capretta 1997; Day 2003; Ge, Donnellan & Wenk 2001; Loeber & Farrington 2000; Mason & Windle 2001; Tremblay & LeMarquand 2001; Wasserman & Seracini 2001). Specific forms of developmental/early intervention include parental training, home visiting, daycare/preschool and home/community programs (Farrington & Welsh 2003). While family and social factors are not readily amenable to policy intervention, there is ample evidence that these programs can be cost effective and reduce offending by about 15 percent (Aos, Miller & Drake 2006; Farrington & Welsh 2003).

Table 1 Risk and protective factors
Risk factors
Child factors Family factors School context Life events Community and cultural factors

Prematurity

Low birth weight

Disability

Prenatal brain damage

Birth injury

Low intelligence

Difficult temperament

Chronic illness

Insecure attachment

Poor problem solving

Beliefs about aggression

Attributions

Poor social skills

Low self esteem

Lack of empathy

Alienation

Hyperactivity/disruptive behaviour

Impulsivity

Parental characteristics

Teenage mothers

Single parents

Psychiatric disorder, especially depression

Substance abuse

Criminality

Antisocial models

Family environment

Family violence and disharmony

Marital discord

Disorganised

Negative interaction/social isolation

Large family size

Father absence

Long-term parental unemployment

Parenting style

Poor supervision and monitoring of child

Discipline style (harsh or inconsistent)

Rejection of child

Abuse

Lack of warmth and affection

Low involvement in child’s activities

Neglect

School failure

Normative beliefs about aggression

Deviant peer group

Bullying

Peer rejection

Poor attachment to school

Inadequate behaviour management

Divorce and family break-up

War or natural disasters

Death of a family member

Socioeconomic disadvantage

Population density and housing conditions

Urban area

Neighbourhood violence and crime

Cultural norms concerning violence as acceptable response to frustration

Media portrayal of violence

Lack of support services

Social or cultural discrimination

Social competence

Social skills

Above average intelligence

Attachment to family

Empathy

Problem solving

Optimism

School attachment

Easy temperament

Internal locus of control

Moral beliefs

Values

Self-related cognitions

Good coping style

Supportive caring parents

Family harmony

More than two years between siblings

Responsibility for chores or required helpfulness

Secure and stable family

Supportive relationship with other adult

Small family size

Strong family norms and morality

Positive school climate

Pro-social peer group

Responsibility and required helpfulness

Sense of belonging/bonding

Opportunities for some success at school and recognition of achievement

School norms about violence

Meeting significant person

Moving to new area

Opportunities at critical turning points or major life transitions

Access to support services

Community networking

Attachment to the community

Participation in church or other community group

Community/cultural norms against violence

A strong cultural identity and ethnic pride

Source: Homel et al. 1999

Crime and place

One approach that may assist with targeting interventions towards individuals on different offender trajectories involves examining the locations where offenders resided when they first had contact with the criminal justice system. Geographic Information System technology is increasingly being recognised as a powerful tool that can be used to enhance organisational decision making, better understand the causes of crime, and target and help assess the impact of crime prevention programs (Anselin et al. 2000; Canter 2000; Hirschfield & Bowers 2001; McEwen & Taxman 1995; Paulsen & Robinson 2004; Taxman & McEwen 1997; Weisburd & McEwen 1997). While the spatial dimensions of data have not previously been explored by criminal careers research, there is reason to believe that offenders may not be randomly distributed geographically.

Studies examining the spatial and temporal distribution of crime are essentially descriptive and typically based on cross-sectional data obtained for short periods of time (Chakravorty & Pelfrey 2000; Eck, Gersh & Taylor 2000; Sherman & Rogan 1995; Weisburd & Green 1994; Weisburd & McEwen 1997). Evidence from these studies indicates that, regardless of the unit of analysis, crime is concentrated in hotspots rather than being randomly distributed (Brantingham & Brantingham 1999; Crow & Bull 1975; Pierce, Spaar & Briggs 1986; Roncek 2000; Sherman, Gartin & Buerger 1989; Weisburd et al. 2004; Weisburd & Green 1994; Weisburd, Maher & Sherman 1992). Sherman, Gartin and Buerger (1989) found that three percent of addresses in their study were responsible for half of the calls to police. Sherman (1995: 36–37) argues that future crime is ‘six times more predictable by the address of the occurrence than by the identity of the offender’. While there is limited research examining how crime is temporally distributed, available evidence suggests that crime hotspots are relatively stable over time (Griffiths & Chaez 2004; Kubrin & Herting 2003; Weisburd et al. 2004).

While there is less evidence about how offenders are spatially distributed, studies conducted in the United States and United Kingdom focused on the journey to crime indicate that most crimes are committed close to the offender’s place of residence. On average, offenders travelled less than five kilometres from their home address to commit offences (Gabor & Gottheil 1984; Phillips 1980; Rhodes & Conly 1981; Townsley & Sidebottom 2010; Wiles & Costello 2000). Young offenders and black offenders have been found to travel less distance to commit offences (Baldwin & Bottoms 1976; Carter & Hill 1979; Davidson 1984; Phillips 1980; Rand 1986; Reiss & Farrington 1991; Rengert & Wasilchick 1985; Reppetto1974). When the locations of crimes and place of residence are aggregated, evidence suggests that most offenders commit crimes within their own neighbourhoods. Gabor and Gottheil (1984) found that three-quarters of a stratified random sample of offences in Ottawa during 1981 were committed by residents rather than out of towners or transients. Pyle (1976) found that 61 percent of those arrested for crimes against the person and 48 percent of those arrested for property crimes in Cleveland over a two year period resided in the same census tract as where the crime occurred. Others have found that the proportion of crimes committed by local residents varied based on the kind of area, with crimes in the outer city more likely to be committed by local residents than crimes in the inner city (Hesseling 1992; Wikstrom & Dolmen 1990).

The notion that offenders are not randomly distributed geographically is also supported by the findings of studies that have adopted an ecological approach. The ecological environments in which individuals are embedded have been found to exert pervasive influences on behaviour independently of individual factors (Kelling 2005; Kubrin & Weitzer 2003; Oberwittler 2004; Triplett, Gainey & Sun 2003). Research that has adopted an ecological approach is based on aggregate level data such as neighbourhoods (Katzman 1981), cities (Harries 1976), or regions (Dienes 1988) and typically involves the use of widely available Census data (Swartz 2000). There is a large body of research indicating that high crime rates are typically concentrated in small geographical areas characterised by structural disadvantage, including low economic status, poverty, segregation, a high proportion of single parent families, residential instability and a large proportion of racial/ethnic minority groups (Bursik 1986; Oberwittler 2004; Sabol, Coulton & Korbin 2004; Shaw & McKay 1969; Silver & Miller 2004; Swartz 2000; Triplett, Gainey & Sun 2003). In their meta-analysis of 214 studies exploring the macro-level predictors of crime, Pratt and Cullen (2005) found that 11 of the 31 predictors had a high independent mean effect size—strength of non-economic institutions, unemployment (length considered), firearm ownership, percent non-white, incarceration effect, collective efficacy, percent black, religion effect, family disruption, poverty and unsupervised local peer groups. Nine of the predictors were reported as having a medium effect—household activity ratio, social support/truism, inequality, racial homogeneity index, urbanism, residential mobility, unemployment (with age restriction), southern effect and arrest ratio.

Findings suggesting that offenders are not randomly distributed geographically hold great promise for the targeting of not only developmental/early intervention programs but also other forms of crime prevention based on geographic location, such as situational crime prevention and community crime prevention. Situational crime prevention focuses on highly specific problems such as types of offending behaviour and the opportunities in specific environments that facilitate offending at particular times and places (Clarke & Felson 1993). The approach identifies 25 techniques that aim to increase the effort, increase the risks, reduce the rewards, reduce provocations or remove excuses (see Table 2). These techniques are based on opportunity theories of crime including rational choice, routine activities and crime pattern theories, which view crime as a product of the interaction between an individual and the characteristics of the setting (Felson & Clarke 1998). While evaluations that have assessed the impact of situational crime prevention on crime are typically short term and methodologically weak, evidence indicates that this approach can result in reductions in crime (Clarke 1997; Eck 2006). Within Australia, this approach has been successfully employed to reduce substance misuse among Indigenous Australians in a range of geographic locations (d’Abbs & Shaw 2008; d’Abbs & Togni 2000; Kennedy 1999; Ray & McFarland 2010; Richards, Rosevear & Gilbert 2011).

Table 2 Twenty-five situational crime prevention techniques

Increase the effort

Increase the risks

Reduce the rewards

Reduce provocations

Remove excuses

1. Target harden:

Steering column locks and immobilisers

Anti-robbery screens

Tamper-proof packaging

6. Extend guardianship:

Take routine precautions—go out in a group at night, leave signs of occupancy, carry phone

‘Cocoon’ neighbourhood watch

11. Conceal targets:

Off-street parking

Gender-neutral phone directories

Unmarked bullion trucks

16. Reduce frustrations and stress:

Efficient queues and polite service

Expanded seating

Soothing music/muted lights

21. Set rules:

Rental agreements

Harassment codes

Hotel registration

2. Control access to facilities:

Entry phones

Electronic card access

Baggage screening

7. Assist natural surveillance:

Improved street lighting

Defensible space design

Support whistleblowers

12. Remove targets:

Removable car radio

Women’s refuges

Pre-paid cards for pay phones

17. Avoid disputes:

Separate enclosures for rival soccer fans

Reduce crowding in pubs

Fixed cab fares

22. Post instructions:

‘No Parking’

‘Private Property’

‘Extinguish camp fires’

3. Screen exits:

Ticket needed for exit

Export documents

Electronic merchandise tags

8. Reduce anonymity:

Taxi driver IDs

‘How’s my driving?’ decals

School uniforms

13. Identify property:

Property marking

Vehicle licensing and parts marking

Cattle branding

18. Reduce emotional arousal:

Controls on violent pornography

Enforce good behaviour on soccer field

Prohibit racial slurs

23. Alert conscience:

Roadside speed display boards

Signatures for customs declarations

‘Shoplifting is stealing’

4. Deflect offenders:

Street closures

Separate bathrooms for women

Disperse pubs

9. Utilise place managers:

CCTV for double decker buses

Two clerks for convenience stores

Reward vigilance

14. Disrupt markets:

Monitor pawn shops

Controls on classified ads

License street vendors

19. Neutralise peer pressure:

‘Idiots drink and drive’

‘It’s OK to say No’

Disperse troublemakers at school

24. Assist compliance:

Easy library checkout

Public lavatories

Litter bins

5. Control tools/weapons:

‘Smart’ guns

Disabling stolen cell phones

Restrict spray paint sales to juveniles

10. Strengthen formal surveillance:

Red light cameras

Burglar alarms

Security guards

15. Deny benefits:

Ink merchandise tags

Graffiti cleaning

Speed humps

20. Discourage imitation:

Rapid repair of vandalism

V-chips in TVs

Censor details of modus operandi

25. Control drugs and alcohol:

Breathalyzers in pubs

Server intervention

Alcohol-free events

Source: Cornish & Clarke 2003: 90

Community crime prevention aims to confront crime at a ‘grass roots’ level in particular local contexts to address those factors within that context that may be causing or maintaining crime (Hope 2001; Kelly & Caputo 2006; Labonte 1997). The factors that ecological studies have found to be related to offending are viewed as contributing to, creating or maintaining offending (Oberwittler 2004). This has led to a range of theories and mechanisms being proposed to explain the relationship between structural disadvantage and crime, such as how specific social processes lead to crime (Oberwittler 2004; Sabol, Coulton & Korbin 2004). Some of the interventions based on this approach are focused on the entire community, while others are focused on the individual. Many aim to facilitate the development of social resources so that communities can effectively address problems (Laverack 2001). Although interventions based on this approach are appealing, few studies have explored their impact on offending or there are conflicting findings. International evidence indicates that mentoring and vocational and educational training programs may be effective for reducing offending (Burghardt et al. 2001; Tolan et al. 2008). There is some evidence suggesting that community economic development programs reduce property crimes and that recreational programs may reduce crime (McCord, Widom & Crowell 2001; Sherman et al. 1997). There is insufficient evidence to conclude that community policing, community mobilisation (such as Neighbourhood Watch) or school after-hours programs reduce crime (Gottfredson, Gottfredson & Weisman 2001; Grinc 1994; Kerley & Benson 2000). While community-based programs operate in many Indigenous communities within Australia, few have been adequately evaluated (see Allard 2011). Available evidence does, however, suggest that night patrols may be an effective way to reduce offending (Blagg 2003; Lui & Blanchard 2001).

One final point that must be considered when focusing on the location of offenders is their mobility. A substantial proportion of the Australian population is mobile and change household address. In 2010, 42 percent of Australians aged over 18 years and who lived in private dwellings had moved within the previous five years, with younger age groups, people renting through private landlords (83%) and the unemployed (62%) more likely to move (ABS 2010). While many of these people may have moved within the same postal area (POA) or Statistical Local Area, this information is not available. Moreover, evidence indicates that individuals are more likely to offend if they have a high number of address changes (Gendreau, Goggin & Little 1996; Hoffman 1994; Worthington, Higgs & Edwards 1999). Therefore, it is essential that research examining where offenders reside explores their mobility. It makes little sense to target government resources and crime prevention resources if hotspots randomly fluctuate over time without intervention (Spelman 1995).

Current study

This project draws on methods and findings from research focused on offender trajectories and crime and place. Findings from trajectory studies indicate that a small proportion of offenders account for a large proportion of offending and costs. While this group of offenders has been retrospectively identified by studies employing trajectory modelling techniques, there is difficulty identifying chronic offenders prospectively. For example, there is no research that has adequately differentiated between identified trajectory groups based on risk and protective factors. Despite this, recent findings indicate that Indigenous Australians are most overrepresented in chronic offender groups. Research focused on crime and place has found that the geographic locations of crime and offenders are not randomly distributed.

Given these findings, the project aimed to assess whether communities could be identified that generated chronic offenders and carried substantial cost burdens associated with offending. If such communities could be identified, they would be ideal locations to target early/developmental crime prevention programs. These programs target potential offenders and aim to move them off of a chronic offender trajectory by addressing risk and protective factors. Evidence indicated that these programs are a cost-effective way of reducing offending for non-Indigenous populations. Communities generating chronic and costly offenders would also be ideal locations to target situational and community crime prevention interventions. These interventions aim to reduce crime by altering the immediate or contextual environment in which crime occurs. In assessing whether communities generate chronic offenders, the project focused on the offenders first recorded residential postal area when they had contact with the criminal justice system but acknowledges the importance of, and examines the extent of, offender residential mobility. There were six research questions addressed by this project:

  • How many distinct offender trajectories can be identified?
  • What are the demographic, offence and criminal justice system event characteristics associated with trajectory group membership?
  • What are the costs of offender trajectories?
  • Are some communities more likely than others to generate chronic offenders?
  • How residentially mobile are chronic offenders?
  • Which communities carry the cost burden of the chronic offenders?