Criminological Highlights Vol. 21, No. 3 - March 2024

This issue of Criminological Highlights addresses the following questions:

1.    Why are Indigenous youths over-represented in Australia’s criminal justice system?

2.    Are judges responsible for the high rate of Indigenous imprisonment in Canada?

3.    Are “streetworker” programs an effective way of reducing gang violence?

4.    Would crime decrease if prisoners didn’t serve their full sentences?

5.    Are programs designed for early release from prison responsible for crime?

6.    Are sex offender registries useful?

7.    Do court-imposed conditions for those on pretrial release keep people from offending?

8.    What needs to be done to improve the effectiveness of community corrections?

Item 1

The over-representation of Indigenous youths in the New South Wales (Australia) justice system is largely a function of the fact that the police treat Indigenous youths differently from non-Indigenous youths. Compared to non-Indigenous youths, Indigenous youths are more likely to be prosecuted rather than cautioned.

Indigenous youths in Australia are much more likely to be imprisoned than are non-Indigenous youths. Although this might appear to be a result of differential judicial decision making, it would be naïve to assume that this difference is solely a function of sentencing decisions (see also Criminological Highlights 21(3)#2). This paper looks at one plausible explanation not involving sentencing: that the police are more likely to charge Indigenous youths than they are non-Indigenous youths.

In New South Wales, Australia, certain youths who are apprehended by the police can be cautioned rather than being charged and sent to court. There is no appeal from such a decision. The consequences of being charged rather than cautioned can, of course, have long term effects. Being charged is likely to increase the likelihood of pretrial detention and being brought to court if the youth is apprehended after this first decision.

The study looked at records of 38,439 youths who had been apprehended by the police in connection with an offence for which they could be eligible for a police caution. In each of the 23 offences that were included in the analyses, Indigenous youths were more likely to be prosecuted than were non-Indigenous youths. More importantly, “after adjusting for age, gender, offence seriousness, area of residence, concurrent offences, prior contact with the criminal justice system and… year, Indigenous status remains a strong predictor of whether a young offender will be prosecuted. The odds of being prosecuted for a caution-eligible offence are 1.83 times higher if the offender is Indigenous than if the offender is non-Indigenous” (p. 263). Indigenous youths of all ages were more likely to be charged than non-Indigenous youths of the same age and this difference was larger as the age of the youths increased (see also Criminological Highlights 18(6)#1).

Unfortunately, the data do not answer the question of why Indigenous youths were more likely than non-Indigenous youths to be charged. One obvious possibility, of course, is that Indigenous youths may be seen as being at higher future risk of offending.  It is also possible that Indigenous youths respond differently to being apprehended and detained by the police. There may also be differences in their background and families. But given that “The Indigenous juvenile detention rate in Australia is currently 20 times that of non-Indigenous juveniles” (p. 269), the findings are of concern no matter what the immediate explanation might be.

Data from other studies show that youths charged are more likely to reoffend and be arrested than those treated informally (Criminological Highlights 14(6)#1). Hence, these results would suggest that Indigenous youths’ offending rates are likely to be disproportionately increased by the fact that the police are likely to charge rather than caution Indigenous youths for relatively minor offense early in their lives.

Conclusion: The findings from this paper demonstrate that Indigenous youths in Australia are more likely than non-Indigenous youths to be dealt with formally by the criminal justice system. The decision to charge or process a youth informally is often not examined carefully, in part because both choices are available for many youths and the decision is not appealable. But given the likelihood that this decision will have long term impacts (on offending and subsequent decisions), it should be given substantial attention.

Reference: Weatherburn, Don and Brendan Thomas (2023). The Influence of Indigenous Status on the Issue of Police Cautions. Journal of Criminology, 56(2-3), 253-277.

Item 2

In 2017-2018, Canada imprisoned its Indigenous people at a rate that is not only dramatically higher than the rate for non-Indigenous Canadians but was marginally higher than the overall imprisonment rate in the quintessential mass imprisonment country – the US – for that period. Data from Ontario suggest that the problem of high rates of Indigenous imprisonment starts long before the sentencing process.

Canada’s Criminal Code recognizes the need to address the high rate of Indigenous imprisonment. Unfortunately, focusing this concern only on the sentencing provisions of the law implies that the problem can be adequately addressed at that stage. This paper suggests that there is a need to address the problem in a much broader fashion.

Imprisonment rates – usually expressed in terms of imprisonment counts (or people in prison on an average day) per 100,000 residents – are not routinely available in Canada for Indigenous people. But piecing together bits of data from various one-time-only sources in 2017-2018, it becomes clear that Canada’s Indigenous people are imprisoned at a rate (677 per 100,000 Indigenous residents of Canada) that is slightly higher than the overall US rate (655 prisoners per 100,000 residents). This rate is much higher than the rate for non-Indigenous Canadians (79 per 100,000).

Canada’s Criminal Code instructs judges to consider non-prison sanctions when appropriate for all those being sentenced. But it instructs judges to give special attention to the circumstances of Aboriginal offenders (s. 718.2(e)). This could reasonably be interpreted as implying that the law reflects a belief that sentencing decisions by judges are largely responsible for the high rate of imprisonment of Canada’s Indigenous people.

Looking at data from Ontario (Canada’s largest province), however, challenges this view. First of all, Indigenous people, under the control of the Ontario correctional authorities, have more extensive histories of contact with the correctional system than do non-Indigenous offenders. They are, for example, more likely to have a history of imprisonment, conditional sentences, and probation than non-Indigenous prisoners. In terms of their most recent offences, when one looks at those who recently completed a prison sentence, Indigenous people are considerably more likely to have served the sentence for a violent offence and are more likely to have a history of probation breaches and other administration of justice breaches. However, their sentences – notwithstanding what looks like a more serious criminal justice history – were shorter than the sentences imposed on non-Indigenous prisoners. 

Conclusion: These findings would seem to imply that to understand why Canada has so many Indigenous people in its prisons, one should look beyond the sentencing process to offending rates, police scrutiny and discretion as well as prosecutorial discretion. This is not a new suggestion: data from Australia (e.g., Criminological Highlights 21(3)#1, 6(3)#4), 20(5)#2) suggest that when explaining high rates of Indigenous imprisonment there (or in explaining variation in imprisonment across states), one has to look beyond variation in sentencing. Similarly, a Canadian paper on the serving of sentences for second degree murder suggested that Canadian judges were less harsh in setting parole ineligibility periods for Indigenous offenders compared to non-Indigenous offenders. But whatever “benefit” judges gave to Indigenous offenders was undercut by more severe treatment by federal correctional and/or parole authorities (Criminological Highlights 20(2)#4): Their parole ineligibility periods were shorter but, as noted by the Auditor General of Canada, Correctional Service Canada is less likely to afford them the timely access to programs that would convince the parole authorities to release them.

Reference: Sprott, Jane B., Cheryl Marie Webster and Anthony N. Doob (2024). Criminal Justice Reform and the Mass Imprisonment of Indigenous People in Canada. In Campbell, Kathryn and Stephanie Wellman Justice, Indigenous Peoples, and Canada. (Routledge).

Item 3

The idea that gang violence can be reduced by having neighbourhood-based “gang outreach workers” or “streetworkers” working directly with gang members in the community is an attractive alternative to traditional criminal justice approaches to community violence since, in theory, it is dealing with the “source” of the problem. The only problem is that research over 60 years has failed to demonstrate positive effects.

Starting in the 19^th century in the US, street gangs began to be seen as natural products of the social conditions in which gang members lived. The characterization of street violence and the explanations for it changed over time such that by the late 1990s, street violence was often described as a “public health issue” (p. 762) requiring social intervention. These approaches contrasted dramatically with the alternative that was popular in the 1970s and 1980s: law enforcement approaches resulting, typically, in incarceration. In contrast to a focus on law enforcement, the idea was that streetworkers would “take the programs to the gangs” (p. 762) and would stop gang violence before it started.

The specific focus of streetworker programs varied across time and place. Early programs focused on new residents of large cities. As time went on, interest developed in addressing the challenges faced in racially segregated areas of cities. Later, the programs were often linked to the War on Poverty in the US. By the end of the 20^th century, some programs focused on keeping youths from joining gangs, as law enforcement approaches became popular  responses to dealing with those already involved in violence. But the idea throughout this history was that streetworkers would work directly with youths in the community.

The explicit goals of these programs also shifted somewhat. The earlier programs focused broadly on delinquency. Later programs tended to focus on gun violence. Hence the type of studies that were carried out on their effectiveness also varied (e.g., from what was happening to targeted individuals or groups to the amount of (gun) violence in a community).

The early studies of streetworker programs make it “clear that such programs do not decrease – and can even increase – delinquency” (p. 771) (see Criminological Highlights 5(4)#1). More recent papers report mixed results. This paper, however, suggests that the more recent papers may overstate the value of these streetworker programs. The current paper presents a detailed analysis of a Boston, Massachusetts, program designed to reduce violence. Interventions involving street outreach, mediation, and the provision of social services with 20 gangs were carried out. Each gang had its own streetworker who also had access to detailed data on what was happening in that gang’s neighbourhood. Most gang members had contact with the streetworker. The gangs receiving these interventions were not chosen randomly. However, comparison gangs were identified and matched with the treatment gangs in a number of independent ways to ensure that the results were not a result of a single flawed approach to identifying a comparable comparison gang.

There was no evidence that the streetworker program used in this intervention had any effect on reducing violence. The results were similar across different methods of analysis. Given that the program was implemented effectively, the failure of the program to reduce violence cannot be attributed to a failure to deliver the program.

Conclusion: Clearly, streetworker programs like the one described in this study cannot be expected to reduce violence in the short term. However, programs like these may provide a mechanism “to reach and engage those left behind by nearly all nonpunitive social institutions” (p. 790). To the extent that streetworker programs can be reoriented to “bring the most harmed and marginalized [residents] into supportive relationships with their communities” (p. 790), there may be hope that they will have a positive impact on the communities in which they operate. It is equally clear, however, that any such program needs to be evaluated carefully during its implementation and operation. Success cannot be assumed.

Reference: Hureau, David M, Anthony A. Braga, Tracey Lloyd & Christopher Winship (2023). Streetwork at the Crossroads: An Evaluation of a Street Gang Outreach Intervention and Holistic Appraisal of the Research Evidence. Criminology, 61, 758-794.

Item 4

Prisoners often do not serve their full prison sentences as imposed by the court. Does the ‘actual’ amount of time an offender spends in prison affect post-release recidivism?  A set of legislative changes in Sweden demonstrates that time in prison and recidivism rates are unrelated.

There are a number of competing theories about the relationship between the amount of time served in prison and subsequent offending. First is the possibility that long sentences interrupt a criminal career. Second, a long sentence might increase the deterrent impact of the sentence and third, a long sentence might lead to increased likelihood of receiving rehabilitative treatment. Alternatively, longer sentences might increase the likelihood of acquiring skills supporting offending. Or prisoners could increase their criminal contacts, and long sentences could raise the likelihood of labelling and stigmatization.

This paper examines what the actual effects are of serving sentences of different lengths (See also Criminological Highlights 21(3)#5). It takes advantage of three changes in Swedish law (in 1983, 1993, and 1999) each of which changed the amount of time that a person, sentenced to a fixed period of time, would actually serve in prison. Each of the three changes affected the likely release date of those being sentenced. The first change decreased the amount of time that would be served; the latter two changes increased it. Because these were abruptly imposed changes, it was plausible to use offenders who were sentenced in the “pre-reform” period as a comparison group.

A total of 3002 offenders who were imprisoned in Sweden for the first time one year before or one year after each of the three reforms were tracked for ten years after their expected release date. Reconviction and reincarceration were both examined. Various controls were included in the analyses including sentence length, age, prior convictions and the frequency of various types of crime, as well as the number of prison days imposed on the offender. A number of different analytic strategies were used in order to ensure that the findings were not associated with a particular form of analysis. Before and after each of the three reforms, the overall sentencing pattern was very similar, suggesting that changes in the release rules did not affect sentencing.

Changes in the rules relating to release from prison resulting in differences in the amount of time spent actually in prison for a given sentence length did not have any consistent impact on the likelihood of being reconvicted for an offence or reincarcerated. “Utilizing large-scale administrative data containing all convictions for Swedish cohorts born after 1957, we find little evidence that increasing or decreasing the length of incarceration has any specific preventive effect on post-release offending” (p. 991).

Conclusion: “From an international point of view, Swedish prisons have one of the highest per inmate expenditures. This is in part due to the small-scale prisons, a low staff-to-inmate ratio, and the extensive rehabilitation programs” (p. 992). Nevertheless, there was no evidence that spending more time in such prisons affected – through rehabilitation or any other mechanism – the likelihood that a person would reoffend or be reincarcerated. “The overall crime-control benefits of increasing incarceration time for first-time incarcerated offenders may be questioned” (p. 993).

Reference: Al Weswasi, Enes, Fredrik Sivertsson, and Olof Bäckman (2023). Does Sentence Length Affect the Risk for Criminal Recidivism? A Quasi-Experimental Study of Three Policy Reforms in Sweden. Journal of Experimental Criminology, 19, 971-999.

Item 5

Reducing the length of time that offenders sentenced to prison actually spend in prison can be done with no effect on rearrest, reconviction, or reincarceration.

Punishment by way of imprisonment is expensive and has been shown to have a number of negative impacts on society. It also has little, if any, effect in reducing crime (see Criminological Highlights “special issues” on The Effects of Imprisonment and Issues Related to Harsh Sentences.) 

This paper examines Oregon’s “Short-term transitional leave” (STTL) program where prisoners, after serving at least 6 months in prison, can be released for up to 90 days (originally 30 days) before the end of their sentences for the purpose of aiding the prisoner to transition to housing, employment, and relationships with their families. The goal of the study was to determine if the program affected rearrest, reconviction or reincarceration within three years of their actual release.  Technical violations were also tracked.   

Only about 22% of prisoners were granted STTL. In order to create a comparison group from the remaining 78% of prisoners, “propensity score matching” was used to match each of 4,898 prisoners who were granted STTL with a prisoner who was not released early but looked very much the same as a prisoner who was granted STTL. In this study, they were matched on 117 variables, including demographic variables, current offence types, various aspects of their incarceration (e.g., disciplinary reports, length of stay), criminal history, and their “criminogenic needs.” Because there were so many prisoners who did not receive STTL, it was possible to create a comparison group that, on these variables, looked very similar to those who received STTL.

Those who received and those who did not receive STTL looked very similar on three critical outcome measures during the three years following release: rearrest (50.6% and 49.0%, respectively), reconviction (40.7% vs. 40%, respectively) and reincarceration (20.8% and 20.1%, respectively). (See also Criminological Highlights 21(3)#4). None of these differences was statistically significant, notwithstanding the very large sample size.  However, there were statistically significantly more technical (supervision) violations in the STTL group than in the comparison group (27.5% and 22.7%, respectively). The difference between the STTL and comparison group in the number of technical violations increased as the STTL times got longer.  It is likely that the reason for the difference between the groups on technical violations was that prisoners released on STTL were subject to more intensive supervision such as more check-ins and home visits by supervisors.

Conclusion: The results of the study suggest that releasing adults from incarceration sooner than their prescribed sentence can be done without creating any increase in offending. The increase in technical (supervisory) violations obviously has no effect on public safety and is, it would appear, under the control of the correctional authorities.  The cost savings from the reduced use of imprisonment compared to community supervision is, in most jurisdictions, considerable.

Reference: Leymon, Mark G., Christopher M. Campbell & Kris Henning (2024). Oregon’s Transitional Leave Program and Recidivism. Criminal Justice and Behavior, 51(1), 43-65.

Item 6

Sex offender registries are an attractive way for legislators to look good by acting as if they are protecting ordinary citizens from a small group of people – sex offenders. The implication is that all sex offenders are especially dangerous, and that “registration” will reduce offending. The problem with this approach is that it is not supported by empirical evidence. After about 10 years of offence-free living, sex offenders are no more likely to commit a sex offence than are people without a record of sex offending.

“Individuals with a history of sexual offending are often perceived as presenting an enduring risk for sexual recidivism” (p.2) leading to special procedures – e.g., sex offender registries – that are justified on the basis that they will protect society from these especially dangerous people. The problem is that the underlying theory – once a sex offender, always a sex offender – is not supported by the evidence (e.g., Criminological Highlights collection Some Recent Research on Sex Offenders). This paper goes one step further than most of the research in this area by comparing the lifetime likelihood that anyone will commit a sex offence to the likelihood that a person convicted of a sex offence will commit another offence.

Not surprisingly, a person who has recently been convicted of a sex offence is more likely to commit a new sex offence than is someone who has never been convicted of such an offence.  But the likelihood that a convicted sex offender will reoffend decreases dramatically over time. However, it never reaches zero. It should be remembered that two other relevant groups also have non-zero probabilities of committing sex offences: (a) those convicted of other offences, not including sex offences, and (b) those who have no criminal record. Each of these groups has a probability greater than zero of committing a sex offence.

This paper looks at sex re-offending rates of those convicted of sexual crimes in British Columbia in 2006 and 2011 as well as the cumulative rate of conviction for males with no criminal records. Looking at the general population of British Columbia males, it appears that about 1.5% of males would be convicted of a sex offence sometime in their lives. Looking at those with a record of at least one non-sexual offence (but no sex offences), we see that “a nonsexual criminal conviction can be considered to increase the likelihood of a future sexual offence by a factor of 2.5 (from 1.5% to 3.8%)” (p. 10). Not surprisingly, however, those who were at one time convicted of at least one sex offence have a considerably higher rate of sexual offence reoffending (about 4.6% after 4.5 years).  However, after being convicted of a sex offence, “most [people with a history of sex offending] would be indistinguishable from the population base rate after 10 years of sexual offence free [living] in the community…. A significant proportion of individuals on sex offender registries in Canada and other countries would be no more likely to sexually reoffend than the general male population.” (p. 12).

Conclusion: “It is difficult to see how having [the wide range of sex offenders currently on sex offender registries] serves any public protection function, despite strong public support for such measures” (p. 12). There are substantial costs for those on the registries (e.g., restrictions which may hinder peaceful reintegration). The imposition of these restrictions is based on the notion that such individuals are highly likely to commit new sex offences, when in fact their risks are not substantially higher than the risk of sex offending in the general population. Neither, however, is zero. Nevertheless, “Within 10 years [of a sex offence conviction] the residual risk of most individuals with a sexual offence history will resemble that of the general population” (p. 14). Long term, or indefinite, restrictions on their lives do not appear to be an effective way of protecting the public.

Reference: Lee, Seung C., Andrew E. Brankley, and R. Karl Hanson (2023). There Is No Such Thing as Zero Risk of Sexual Offending. Canadian Journal of Criminology & Criminal Justice, 65, 1-31.

Item 7

Adding conditions to an accused person’s pretrial release after that person was charged with a federal offence in the US “generally had no significant relationship with reductions in the likelihood of pretrial crime or missed court appearances” (p. 1868-9).

The decision to release or detain an accused person in custody after they are charged is typically supposed to reflect the risk that they will not appear in court as required or the risk that they will commit a new offence. Some people are obviously more likely to be good candidates for release than others. The question that this paper addresses is whether adding conditions of release affects the likelihood – above and beyond their risk score – that the accused will fail to appear in court as required and/or commit a new offence.

Conditions imposed on those released prior to trial often bear no relationship to the offence (Criminological Highlights 13(5)#5), are experienced as punishment (16(6)#4), can easily lead to new charges of failure to comply with the conditions of release (12(5)#3), and do not reduce crime (15(3)#1).

In this study the re-offending rates of 223,260 people charged with US federal offences and released into the community were examined. Accused people’s “risk” of violating the conditions of release were assessed with a scale developed for this purpose. The scale had modest accuracy in predicting bail violations. The effect of the number of conditions imposed on an accused person was examined independent of their score on the “risk scale”. Other factors such as age, sex, race/ethnicity, charges, etc., were controlled statistically.

Overall, judicial officers making decisions on bail imposed an average of nine special conditions on each accused person. Not surprisingly, these judicial officers seemed to work on the assumption that increasing the number of conditions would reduce risk: More conditions were imposed when risk was assessed to be moderate or high compared to cases where risk was assessed to be low. Equally unsurprising was the fact that many conditions (an average of 10 or more) were likely to be imposed on serious offences (sex offences, drugs, weapons/firearms offences) than when the most serious charge was a property or technical/ public order offence (an average of 4.8 – 7.5 conditions).

The effect of the number of conditions – above and beyond the ‘risk score’ – on three measures was examined: pretrial arrest for any offence, failure to appear, and the pretrial revocation of release. In all cases, the accused person’s risk score predicted, to some extent, the outcome: those with high-risk scores were more likely, not surprisingly, to have poor outcomes. But the impact of conditions – above and beyond the risk score – was negligible. When there was a small effect – for example in the case of pretrial revocation – it would appear that the number of conditions was associated with increased odds of revocation. This isn’t terribly surprising, given that adding conditions to a release order provides more opportunities for revocation.

Conclusion: The findings make it clear that adding conditions to a pretrial release order can increase slightly the probability of revocations for technical violations but generally has no effect on rearrests for a new offence or failure to appear (p. 1866). From an efficiency and fairness perspective, then, it would appear that conditions should normally be imposed very selectively and only in those circumstances where there are compelling reasons for the condition to be imposed on the particular suspect before the court.

Reference: Cohen, Thomas H. and William Hicks, Jr. (2023) The Imposition of Pretrial Conditions on Released Federal Defendants. Criminal Justice and Behavior, 50(12), 1852-1873.

Item 8

Training programs for community correctional staff can – but do not always – affect the nature of the interaction between those serving sentences and those supervising them. In addition, effective programs can – but do not always – reduce reoffending. For the training of community correctional staff to be effective, it needs to be intensive, continuing, and monitored.

Criminal courts in Canada are much more likely to impose probation than they are to impose prison sentences. In Ontario, for example, custody was the most severe sentence in 26% of sentences handed down; probation was the most severe component of the sentences in 66% of cases. In addition, of course, many sentences include probation after a relatively short sentence of imprisonment. Similarly, in the US, more people are serving community sentences than are currently incarcerated (Paper#1-587).

A critical issue, then, is what community correctional programs should consist of. These two papers suggest that probation officers’ supervision time is best spent targeting medium- and high-risk probationers’ criminogenic needs, especially pro-criminal attitudes. The main goal should be “to help clients to replace their pro-criminal thoughts with prosocial thoughts and consequently act more pro-socially” (#2-p. 43).

These two review papers come to similar conclusions. Training community corrections officers to focus their interactions with their clients on correctional principles can be effective. There is little point in focusing on those with low risk to reoffend. Similarly, it is important “that the goals of treatment should be to modify the criminogenic needs or dynamic risk factors” and to focus on cognitive-behavioural interventions (#2-p.40).  This may sound straightforward, but it means that communities will occasionally experience reoffending by a low-risk client who has not received much attention. It also means that there should be a move away from rewarding community corrections officers for surveillance and control since the value of such efforts appears to be minimal. However, the shift to a more structured interaction between a community correctional officer and client is neither easy nor, necessarily, long lasting: “Implementation should not proceed until a monitoring and evaluation system is established” (#2-p.52). 

Given the problems of implementation described in these papers (see also Criminological Highlights 19(4)#2) and the relatively low level of correctional assets invested in community correctional programs, it is not surprising that the impact of community programs on client recidivism is often non-existent or unimpressive. Looking across evaluation research involving high quality studies, there was no significant effect on rearrest, reconviction, or technical violations (#2, p. 601-602), though there was some indication of effectiveness in some (but not all) studies. The focus in these studies is obviously on correctional issues, rather than on other ways in which the correctional officer might be able to help the client.

Conclusion: It is clear that training of community corrections officers focused on risk-need-responsivity principles can affect the interaction between the client and the correctional officer. However, it is not the case that a simple one-time training program can be effective at reducing reoffending in the long term. Constant monitoring and evaluation appear to be necessary to ensure that community correctional resources are used effectively.

References: Labrecque, Ryan M., Jill Viglione, and Michael Caudy (2023). The Impact of Community Supervision Officer Training Programs on Officer and Client Outcomes: A Systematic Review and Meta-Analysis. Justice Quarterly, 40, 587-611 (#1). Bonta, James (2023). Training Community Supervision Officers in the Risk-Need-Responsivity Model of Offender Rehabilitation: A Review and Implications. Journal of Offender Rehabilitation, 62, 39-58 (#2).

This issue of Criminological Highlights was prepared by Anthony Doob, Rosemary Gartner, Maria Jung, Tyler King, Audrey Macklin, Jane Sprott, and Danielle Van Wagner.