Criminological Highlights Vol. 23, No. 1 - May 2026

This issue of Criminological Highlights addresses the following questions:

1. Should people who are released from prison after serving short sentences be supervised in the community?
2. Why is the amount of unstructured spare time important in understanding crime?
3. Why have rates of youth crime decreased in many countries?
4. How can city maintenance policies reduce nighttime crime?
5. Should police officers be routinely stationed in schools?
6. How are risk assessments used by judges making pretrial detention decisions?
7. Does the manner in which police officers treat young citizens affect their views of police in the same way across all groups?
8. Are White, Black, and Hispanic citizens affected in the same way by procedural justice training given to police officers?

Item 1

The supervision of those released from prison sentences of less than a year appears not to reduce reoffending either during the supervision period or after the supervision period has ended.

Offenders in many countries who are being supervised in the community often first served a period of time in prison. This paper compares the frequency of reoffending in the community for those who received supervision after being released from prison to those who did not get supervised after release.

The focus of this study is on those sentenced to one year or less in a Dutch prison. Some (42%) are supervised in the community after they are released; some are not. Supervision in the Netherlands is more likely for violent offences and is less likely for theft and other property offences. Supervision is also more likely to be imposed if the sentence is longer or if the offender had been supervised in the past. Hence, in attempting to estimate the effect of supervision, it is necessary to find people who receive mandatory supervision who are equivalent to those who do not receive supervision. In this study, a matching system was used to find people whose records looked very similar, but where one was supervised after release and one was not. This is important to keep in mind because the results are limited to those where a reasonable match could be found. For the group that was supervised after release from prison, 23% could not be matched. The goal of the study was “to compare supervised persons to otherwise similar persons who were not supervised” (p. 446). It means that extreme cases who could not be matched (e.g., those being supervised who had served prison sentences close to one year who differed in important ways from the rest of the population) were not part of the sample.

The results were quite straightforward. During the supervision period (or equivalent period for those not being supervised), those who were being supervised were just as likely to reoffend as were those who were not being supervised. If they did reoffend, the rate at which the two groups reoffended did not vary. The timing of the former prisoners’ contact with the supervising officer after they were released from prison varied somewhat, but the timing of this first face-to-face contact did not seem to predict anything.

Similar results were found when the researchers examined reoffending after the formal supervision period had ended. Hence there was no evidence that the supervision of those who were released into the community after serving prison sentences of up to a year had any effect on the likelihood that the former prisoner would commit an offence.

Conclusion: The fact that supervision of those released from short prison sentences does not appear to reduce reoffending is not a terribly surprising finding. “Prior research suggests that supervision programs that are based on surveillance do not reduce crime… Although in the Netherlands supervision was supposed to be focused on both surveillance and support, the emphasis, during the time of this study, might have been too much on surveillance… [Most prisoners who were interviewed] perceived supervision as predominantly aimed at monitoring, while in their view little effort was given to offer assistance” (p. 455).

Reference: Verweij, Suzan and 5 others (2026). Supervision Following Release from Prison and Reoffending During and After the Supervision Period. Crime & Delinquency, 72(2), 436-465.

Item 3

Crime rates by young people in many countries, including Canada, have decreased considerably in the past 20-30 years. Two factors appear to be associated with the drop: declines in unstructured socialization with peers and declines in alcohol consumption.

Official and self-report studies of offending by youths show widespread decreases in youth crime in recent decades. For example, in Canada, police data for youths (charged and not-charged per 100,000 youths) declined between 2003 (the first year of Canada’s current youth justice law) and 2024. For all offences (charged and not-charged combined) there was a 65% decline; for violent offences the decline was 29%.

Typically, the drop has been greater for property crimes than for violent crimes. But in addition, the decline in offending rates appears to be much greater for youths in the highest offending ages (late teenage years). Not surprisingly, given that boys are more likely to be involved in violence than girls, the self-report data on being involved in a physical fight show much more consistent declines across countries for boys than for girls. Crime by young people can decline because fewer youths are involved in any crime (prevalence), or because those who are involved are doing less of it (incidence) or both. There is some evidence from some countries that the decrease is primarily from decreases in prevalence.

The data are impressively consistent; and although there is some variation (e.g., across countries and groups within countries), the general decline in youth crime “calls for broad and parsimonious explanations” (p. 182) for the change.

Many studies (e.g., Criminological Highlights, this issue, Item 2) have suggested that there have been substantial declines in the amount of unstructured and unsupervised face-to-face interactions of youths with their friends. Research has suggested that unstructured socializing – e.g., hanging around without authority figures or other forms of social control – is strongly associated with delinquency. One US study, for example, found that “unstructured socializing with friends declined… by 30% between 1999 and 2017” (p. 191).

A second factor that appears to have changed in countries where data are available is alcohol use. In studies both in the US and in Sweden, for example, the decline in alcohol use by young people accounts for substantial proportions of the decline in offending. In contrast, two other plausible factors – changes in parental and school bonds – accounted for very little of the change in offending. In other studies that implicate alcohol and unstructured socializing as explanations for the change in offending by youths, it was found that attitude measures (related, for example, to the neutralization of delinquent behaviour) did not contribute to the understanding of change in offending in recent years.

Conclusion: “The core takeaway [from this review] is the paramount role of changing routine activities on youth crime trends. The declining relevance of unstructured socializing and alcohol consumption in adolescents’ daily lives consistently turned out to be by far the most important factors explaining the decline in youth crime…. These findings clearly boost routine activity and situational theories of crime causation” (p. 196) as explanations for the decline in youth crime. On the other hand, “If unstructured socializing and alcohol consumption are proximal causes of crime, the question remains of what more distal factors influenced adolescents to change their daily life routines….” (p. 179).

Reference: Oberwittler, Dietrich and Robert Svensson (2025). The International Youth Crime Drop: Evidence and Explanations. Crime and Justice (Michael Tonry, editor), 54, 153-216.

Item 4

Improving the lighting on city streets reduced nighttime street crimes, generally, and gun violence, in particular, in a large US city.

Previous research has suggested that improved lighting on city streets can increase public safety and reduce the fear of going out at night. Typically, these studies have focused on the impact of improved lighting in a small number of discrete locations. The concern about such improvements in lighting in small areas is that crime may simply move to traditionally poorly lit areas of the city. In contrast, this study examines the City of Philadelphia’s efforts to upgrade all city streetlights.

Between August 2023 and May 2024, 13,275 city blocks (about one-third of the city) in Philadelphia had their streetlights upgraded to energy efficient and better-quality LED lights. Data were analyzed at two levels: street segments and hexagons (n=720) each covering roughly half a square kilometer. For each hexagon, the number of crimes and the number of updated streetlight fixtures were recorded for each month.  Because the city initially concentrated its installation of upgraded lights in those areas with high crimes, it was important to look at the change in crime that was associated with the installation of the new lights.

The findings are straightforward: “Lighting upgrades are associated with a significant reduction in overall crime [a 5% decrease] as well as gun crimes [a 17% decrease], violent crimes [4% decrease], and property crimes [a 7% decrease]” (p. 43). There was no impact of streetlighting upgrades on financial crimes (typically committed indoors) or “nuisance crimes” (which tend to be police-discovered).  A separate analysis demonstrated that the decrease in crime associated with the upgraded lighting system was “primarily driven by crimes occurring outdoors” (p. 44).

“The lighting upgrades resulted in larger crime reduction during nighttime hours compared to daytime hours with overall crime declining by 8% at night versus 4% during the day” (p. 45).

Observations made during the period when lights were being upgraded noted that there was a fair number of non-functioning streetlights on blocks with older lights.  Areas with the upgraded streetlights had “better visibility and no obvious dark areas” (p. 50). Residents reported that “they try to stay on streets with good lighting when travelling at night and generally avoid going outside after dark if they can” (p. 50).

“A key advantage of investments like enhanced street lighting is that they do not require charismatic leadership, heroic mid-level managers, or rare expertise for installation or maintenance. Additionally, [the] interviews with community members suggest that residents view upgraded lighting as a positive sign of city investment, even if they did not think lighting on its own was a sufficient crime prevention approach” (p. 53).

Conclusion: This paper demonstrates that the installation of good street lighting on crime-prone streets can reduce various forms of serious crime (e.g., violence and gun crimes). The period of time in which crime was monitored in this study was quite short (10 months). But, at least for this period – immediately after the installation – the crime reduction effects appear to be reliable.

Reference: MacDonald, John M. and 6 others (2026). Can Enhanced Street Lighting Improve Public Safety at Scale? Criminology & Public Policy, 25¸ 31-62.

Item 5

Stationing sworn police officers in schools does not appear to contribute to the physical safety of those in the school.

The practice of stationing law enforcement officers in schools has been growing in many countries. The primary purpose is to maintain school safety. Often it is high profile gun violence that leads to demands to have police officers in the school. However, evidence that students and staff are safer in schools in which police officers are stationed is seldom collected.

Advocates of the use of school-based law enforcement (SBLE) officers often suggest that there are two, quite different, benefits of the practice: (1) that the presence of law enforcement officers will reduce crime and/or increase safety, and (2) that law enforcement presence will increase feelings of safety. This paper focuses, primarily, on the first suggestion.

There are many challenges inherent in assessing this research. There is no general agreement about what SBLE consists of (e.g., the density of police presence or what they actually do if not responding to an emergency), nor is there a standard training program or agreement on who (e.g., the school or the law enforcement agency) directs the activities of the officers.

For this systematic review of the research literature on the impact of SBLE, no true (random assignment) experiments were found. Instead, the authors were forced to rely on studies where there were measures obtained in schools during periods when there was and was not a SBLE program and studies in which schools with SBLE were compared to schools that never had such programs. A total of 32 studies were found.

Twenty-seven studies that included estimates of the impact of SBLE on crime and behaviour were located and examined. The data suggest that schools with SBLE had more problems with crime and behaviour as well as higher rates of discipline than schools without SBLE. However, when looking at individual student reports, there was no effect of SBLE on crime and behaviour. There was also a very small effect whereby students in schools with SBLE experienced a slightly larger amount of discipline. There were no effects of SBLE on measures of violence, weapon use or possession, substance use or possession, or criminal justice contact. Similarly, there were no effects on students’ perceptions of the school.

There was some indication that schools with SBLE were perceived as being safer, but this was only the case when “schools” were the unit of analysis. The data for individual students showed no difference between students in schools with and without SBLE. There was no effect found on learning outcomes or school attendance.

Conclusion: Although there is some indication that students in schools in which police officers are stationed attribute feelings of safety to the presence of these officers (see also Criminological Highlights, 21(2)#3), this systematic review of the literature suggests that the impact on crime, safety, and related matters is varied across studies and, overall, does not appear to be favourable. Two conclusions appear to be justified. First, we cannot assume that the placement of police officers in schools will reduce crime and disorder. Second, if the placement of police officers in schools is to occur, it would appear to be important to do it in such a way that the various impacts can be evaluated.

Reference: Fisher, Benjamin and 6 others (2023). School-based Law Enforcement Strategies to Reduce Crime, Increase Perceptions of Safety, and Improve Learning Outcomes in Primary and Secondary Schools: A Systematic Review. Campbell Systematic Reviews, 2023 (December).

Item 6

In making decisions related to pretrial detention, judges sometimes have access to formal risk assessments of an accused person. Judges selectively invoked these risk scores when they found it useful to do so to justify their decisions. The selective “uses of risk scores effectively insulated judges from institutional critique and legitimated their punitive sanctions” (p 166).

Decisions about pretrial release are inherently based on risk assessment: largely the likelihood that a person will not appear as required in court and the likelihood that the accused will be charged with another crime. Formal risk assessments that have developed in recent years appear to provide “objective” evidence of the likelihood that either event will occur.

The problem in using a risk assessment tool is simple: A risk is usually expressed as a score on a continuum (from very low to very high risk that the accused will misbehave) but the decision is essentially binary (release or not). Judicial officials make the final decision; the ‘risk score’ is employed (or not) at the discretion of the judge. This study is based on interviews with judges in four US states who regularly preside over pretrial hearings and is supplemented by interviews with prosecutors and public defenders as well as some court observations. 

In many instances, the value of the risk assessment was simple: it provided quick guidance in sorting out opposing views from the two sides, though judges tended to find it most useful when the risk was at one or the other extreme: the risk assessment tool provided justification for an obvious decision. Nevertheless, in many instances judges expressed the view that these objective risk assessment tools did not give sufficient weight to certain details of specific cases that they saw as relevant to their decision. 

At the same time, judges sometimes indicated that they invoked the risk assessment score to justify a controversial decision. As one judge put it, “Sometimes we’ll use it as a justification to do something that a victim doesn’t want us to do” (p. 165). In addition, these “strategic uses of risk scores effectively insulated judges from institutional critique and legitimated their punitive sanctions” (p. 166). 

More generally “risk scores [are sometimes] introduced by judges who invoke them when they mitigate institutional tensions and dismiss them as invalid when they do not” (p. 166). Risk assessments, then, act as “culturally authoritative tools that instill legitimacy in organizational decision-making processes” (p. 167).  “Judges strategically reference risk scores in pretrial hearings to manage tensions within and among institutional logics that govern their decisions” (p. 167)

Conclusion: “When selectively invoked, risk scores serve as legitimating resources that judges can harness to stymie political blowback for decisions that result in unfavourable outcomes (i.e., when defendants commit new crimes or fail to reappear in court)” (p. 166). “Risk scores may be better conceptualized as resources that actors situationally use to navigate institutional tensions and legitimate their decisions. This is especially likely in “unsettled” contexts such as pretrial hearings where judges have limited access to decision-making resources” (p. 167). “Today’s risk assessments are considered especially credible because they use algorithms that process massive amounts of data…. Yet they are also impaired by measurement errors that call their legitimacy into question” (p. 167).

Esthappan, Sino (2026). Assessing the Risks of Risk Assessments: Institutional Tensions and Data Driven Judicial Decision-Making in U.S. Pretrial Hearings. Social Problems, 73, 156-170.

Item 7

The views that adolescents and young adults hold of their local police are important. But the way in which assessments of the police change for those who have been involved in crime is not straightforward and varies across racial/ethnic groups.

“Police contacts are ‘teachable moments’ that give adolescents the opportunity to directly appraise the fairness and legitimacy of law enforcement” (p. 29). It seems likely that individuals continuously revise their views of the police based on new information and experiences. When looking at assessments of whether the police act in a procedurally just manner, however, there is some evidence that negative experiences may have more effect than positive experiences.

This paper examines data from a longitudinal study of youths in three US cities. All youths had been found guilty of at least one serious offence and, therefore, are more likely than most youths to have had direct contact with police. Respondents represented three racial/ethnic groups: Whites, Blacks, and Hispanics. They were interviewed regularly over a period of 7 years starting when they were 14-17 years old. A scale focusing on views of police legitimacy (e.g., respect for the police, police honesty) was used for the main assessment. In each interview, youths were questioned about their contact with the police and were divided into four groups based on their recent experiences: Not picked up by the police, and picked up by the police and disrespected, treated neutrally, or respected. Various controls were also included such as the youth’s gang involvement, time on the streets, and peer delinquency.

Pooling across all three racial/ethnic groups, being picked up and being treated in a manner described in neutral terms reduced slightly the ratings of police legitimacy. Being treated in a manner described as being respectful increased slightly (but significantly) the ratings of police legitimacy. The largest effect, however, was for those who indicated that they had been disrespected. For these respondents, there was a large decrease in the rated legitimacy of the police.

White and Hispanic respondents were quite similar to the overall sample. Black respondents, however, were quite different. Like the others, Black youths who indicated that they had been picked up by the police and treated in a disrespectful manner showed a large decrease in their ratings of police legitimacy.  However, unlike White and Hispanic respondents, Black youths who had been picked up by the police and treated in a respectful manner were no different from those who had not had not been picked up by the police. In other words, being treated in a disrespectful manner reduces views of police legitimacy for Black youths, but being treated in a manner described by the respondent as respectful did not increase ratings of police legitimacy for this group of young people.

More generally, looking across all groups, the size of the effect of “disrespectful interactions” is more than three times greater in magnitude than that of respectful interactions when compared with not being picked up by the police. What these findings suggest, then, is that disrespect is a particularly strong signal that hinders police legitimacy, even in adversarial police-initiated encounters under study here” (p. 48). It would seem that a respectful police encounter with a Black young person cannot overcome “the distrust engendered by generational systemic inequities that have been personally experienced” (p. 50).

Conclusion: For White and Hispanic, but not Black youths, positive experiences with the police create small improvements in the assessments of police legitimacy, but for all youths, negative interactions with the police reduce ratings of police legitimacy. It would appear that to improve the views of police legitimacy of those starting off with quite negative views “bold and wide-reaching reforms” (p. 53) may be necessary.

Reference: Thompson, Andrew, Theodore Wilson, and Brandon Behlendorf (2026). Is Updating Racialized? Differential Effects of (Dis)respectful Police Contact on Perceptions of Police Legitimacy. Criminology, 64, 28-58.

Item 8

Police training aimed at ensuring that residents whom police come in contact with are treated more fairly and respectfully in three American cities had comparable favourable outcomes for White, Black, Hispanic and other groups of citizens in these cities.

A previous study (Criminological Highlights 22(1)#4) suggested that police interactions with citizens in high crime areas of a city can be improved through intensive police training in procedural justice. What was not examined in the original study was whether this finding held for various racial and ethnic groups.

This paper, using the data from the earlier study, examined police interactions with citizens. The quality of these interactions was coded by an independent observer on a number of separate dimensions related to procedural justice including the degree to which the interaction reflected fair and respectful treatment of citizens by the police officers (voice, neutrality, dignity, respect). An overall procedural justice measure was created for each interaction. 

A critical feature of the study was that before the period in which observations were made, police officers were assigned at random to one of two treatments: Some officers received a 5-day, 40 hour-long procedural justice training program. Other officers received no special new training. Hence the study was focused on the effect of this intensive treatment. This paper explored whether the effect was the same for different racial/ethnic groups. The interactions were observed by researchers and in all but 7 of the 508 interactions that were observed, the race of the citizen was estimated by this researcher.

The police officers who received procedural justice training had more interactions with citizens than did officers who had no special training. And officers who received procedural justice training had proportionately fewer interactions with Hispanic/Latino citizens.

Looking first at the overall ratings of the interactions between the citizens and the officers who did not receive the special training, it was found that for the four groups of citizens (White, Black, Hispanic, and other), there were some differences in the nature of the interactions: Interactions between officers and White and Other citizens were given more favourable ratings by the researcher than were interactions with Black and Hispanic citizens.

However, for all four groups, interactions with officers who had received the intensive procedural justice training were rated more favourably than were interactions with “control” officers who did not receive special training. In addition, looking at the behaviour of officers, those who received procedural justice training were less likely to arrest citizens than were control officers. Similarly, disrespectful treatment was less likely to be seen with officers who had received procedural justice training than those who observed control officers.

Conclusion: The data show that there was no evidence “that procedural justice training improves police-citizen interactions more for one racial/ethnic group than another” (p. 216). A more pessimistic way of expressing the results would be that the “study does not provide significant evidence of relatively greater improvements in treatment of non-Whites as a result of the procedural justice training” (p. 216). In other words, though all groups seem to benefit from providing this special training to police officers, the training “does little to undo the damage and trauma caused by historical and contemporary inequities in police by race/ethnicity” (p. 216).

Reference: Weisburd, David, Cody W. Telep, and Kiseong Kuen (2026). Does a Rising Tide Lift All Boats? Evidence from a Multicity Randomized Trial of Procedural Justice in Hot Spots Policing. Criminology, 64, 207-218.  

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