Criminological Highlights Vol. 23, No. 3 - October 2026

1. How do the pretrial “performances” of accused people affect the outcome of criminal cases?

2. How are “sureties”, who supervise pretrial release of accused people in Canada, almost inevitably in a role conflict?

3. How are Canada’s rules on the treatment of Indigenous people in the bail process inadequate?

4. Can diversion programs for those charged with minor offences reduce subsequent offending?

5. How can the public’s lack of knowledge about crime and the justice system lead to inadequate development of justice policy?

6. Do safe drug consumption sites lead to increased crime?

7. Are dark-skinned Asian residents of the US more likely to be stopped and questioned by the police than lighter-skinned Asians?

8. Does being arrested increase the likelihood that people will believe that they will be arrested again if they commit an offence?

Item 1



The “pretrial performance” of accused people in the criminal justice system affects the way in which prosecutors develop plea offers. “The practice of demanding and evaluating pretrial performances in plea bargaining can widen the net of the criminal justice system and exacerbate existing inequalities” (p. 259).

Previous research (see Criminological Highlights special issue on bail and pretrial detention) has demonstrated that there are harmful effects of pretrial detention. This paper focuses on the impact of defendants’ informal behaviour prior to trial on prosecutors’ plea offers. The findings suggest an additional way in which pretrial detention disadvantages accused people: those who are detained prior to trial, along with accused people who are disadvantaged in other ways, are less likely to receive attractive plea offers.

The study is based on data from an ethnographic study of prosecutorial discretion and decision making in two counties in a US midwestern state. Prosecutors’ decisions concerning plea offers were based on four types of pretrial performances on the part of accused people: (1) the accused person’s compliance with pretrial conditions that had been imposed on them; (2) whether the accused took steps to address the direct harm of the offence itself (e.g., apologies, restitution); (3) the accused’s engagement in rehabilitative programs; (4) the accused’s general demeanor and attitudes toward the police, prosecutors, and the court.

In some cases, prosecutors require certain behaviours in exchange for a more favourable sentence reduction or the dismissal of charges. In other cases, the accused’s behaviour is simply evaluated without explicit demands from the prosecutor. Defence lawyers often learn what types of performances prosecutors are looking for and may encourage their clients to behave accordingly. However, “prosecutors’ demands and expectations are often hidden from view, occurring behind closed doors during the plea bargaining process” (p. 258). In addition, different courtroom groups are likely to develop “distinct norms around which pretrial performances warrant a better plea offer” (p. 258).

There are, of course, two ways in which such approaches to plea bargains can disadvantage accused people. In the first place, this approach to plea bargaining places responsibility on the accused to know what kinds of actions are likely to be of benefit in determining the outcome of the case. “Ultimately, this practice of evaluating pretrial performances transforms the role of the prosecutor from one of prosecution to one of managing and rehabilitating defendants who are legally presumed innocent” (p. 259). But in addition, “Existing research provides ample evidence that pretrial detention has deleterious consequences for criminal outcomes and reinforces social inequalities….  A defendant’s inability to demonstrate governability through pretrial performances [because they are detained] may constitute an additional pathway through which pretrial detention produces worse outcomes” (p. 259).

Conclusion: “Ultimately, this prosecutorial practice of demanding and evaluating performances pretrial ignores the structural barriers that many people entangled in the criminal legal system face and the lack of adequate resources in the community (e.g., long waitlists….) to meet performance demands. It can, consequently, exacerbate inequalities as more privileged defendants have greater resources to meet the demands of prosecutors…. More disadvantaged defendants [including those who are detained prior to trial] may face harsher punishment [than those with more resources] as a consequence of these practices” (p. 260).

Reference: Packard, Chiara Clio (2026). Evaluating Performances Pretrial: The Role of Defendant Performances in Prosecutors’ Plea Bargaining Decisions. Criminology & Public Policy, 25, 241-263.

Item 2

Courts often require people who are released on bail into the community to have a surety (a person who agrees to ‘supervise’ the accused person and who agrees to forfeit a fixed amount of money if the accused does not do what is required). They “occupy a conflicting role and are often overwhelmed with trying to balance their responsibilities to the accused with their obligations to the court” (p. 168).

In Ontario, Canada, when an accused is released on bail by a judicial official, a surety – typically a friend or relative of the accused – often is required. The requirement of a surety is meant to demonstrate that the accused has ties to the community. But in addition, the surety takes on the responsibility of supervising the accused person and enforcing the conditions of release imposed by the court. Hence they are expected to show higher allegiance to the court than to the accused they are associated with.

In effect, then, although they may have a close personal relationship to the accused person they are supervising, sureties are acting as “court-enlisted bail supervisors or third-party police” (p. 151). In addition, sureties typically pledge to pay the court a fixed amount of money if the accused person violates a condition of their release or fails to appear in court as required. In an Ontario study, sureties were required to enforce an average of 7.8 conditions and to pledge over $1000 if the accused breached conditions of release.

In order to understand the complexities of the role of sureties, 32 people who had acted as sureties in Ontario were interviewed for this study. Those interviewed identified a number of challenges and problems with the role they had agreed to take. Many sureties – notably those who were supervising close relatives – felt that they had no choice but to agree to act as a surety, in part because they saw the alternative – detention until trial – as something they couldn’t be responsible for. At the same time, most participants reported damage to their relationship with the accused as a result of being a surety. Their obligations as a surety (e.g., going to court with the accused) often interfered with ‘normal’ obligations like work. Many indicated that they resented both the situation they had agreed to and the accused.

Respondents reported that they found themselves worrying about the responsibilities they had taken on, and felt stressed by the burden of supervising an adult friend or family member. About half of the respondents indicated that they would never do it again, in part because it involved too much responsibility and stress. Some noted they thought that “release should be the accused person’s own responsibility, without surety supervision” and that “the use of familial bonds for surety supervision may strain those connections, leaving accused people with less community support” (p. 164). “Acting as a surety is [experienced as being] morally conflicting and a significant impediment to the surety’s personal life” (p. 165).

Conclusion: There is no doubt that acting as a surety puts a stress on the relationship between the accused and the surety. In effect, then, the surety is receiving a punishment imposed by the court. There are alternative approaches to community supervision. If there is evidence that the accused needs supervision (and perhaps support such as housing), bail supervision programs (operated by the government or independent organizations) are sometimes used, although there are concerns that such organizations may be unnecessarily selective in whom they will supervise. What is clear from this paper, however, is that “Sureties occupy a conflicting role and are often overwhelmed with trying to balance their responsibilities to the accused with their obligations to the court. A perceived lack of appreciation from both the courts and the accused challenges this tenuous and fragile balance” (p. 168).

Reference: Myers, Nicole Marie and J. McDermott (2026). Breaking Bonds: Rupturing Familial and Social Networks Through Supervisory Bail Conditions. Canadian Journal of Law and Society, 41(1), 149-171.

Item 3

Rules that require courts to consider the special challenges faced by Indigenous people in the bail process are insufficient to address the problem of the over-representation of Indigenous people in pretrial detention in Canada.

In 1996, Canada’s Criminal Code was amended to include a provision requiring judges to give “particular attention to the circumstances of Aboriginal offenders” (Criminal Code S718.2(e)) when implementing a provision giving priority at sentencing to non-custodial sentences. More recently, judicial decisions have made it clear that these same principles should apply when considering whether an Indigenous person should be detained prior to trial.

Although roughly 5% of Canada’s population are Indigenous people, 32% of those admitted to pretrial custody are Indigenous people. This has increased in recent years. In Ontario, for example, in 2023, 19% of those admitted to remand custody were Indigenous people, compared to 13% in 2013 and 9% in 2003.

In this study, interviews with 46 Ontario criminal defence lawyers were carried out in an attempt to understand the continued over-representation of Indigenous people admitted to remand custody. Most of those who were interviewed reported that the special concerns regarding Indigenous people were “imposed superficially and with limited impact” (p. 96). It was pointed out, for example, that Indigenous people were less likely than others to have access to sureties (people who were deemed by the court to be appropriate to take on the responsibility of supervising the accused and who were able and willing to forfeit money if they failed in their supervision). Similarly, giving special attention to the history and current circumstances of Indigenous people did not substitute for not having a strong and persuasive plan for supervising an accused person if they were released. Defence counsel noted that legal aid funding does not adequately address the time required to develop a strong release plan for a typical Indigenous person.

Indigenous lawyers reported that the reality of the lives of most Indigenous accused people was such that it was impossible to convey to the court the special circumstances of their clients and how, given these special circumstances (e.g., inadequate housing, family members with criminal records, lack of employment opportunities), release of the accused person was appropriate. 

The general principle, generalized from Canadian sentencing laws, that the circumstances and historical treatment of Indigenous people should be taken into account, does not appear to translate well when applied to the primary concerns related to the choice between pretrial detention and release on bail: i.e., concerns about whether the accused will appear as required in court and desist from offending. Indeed, outlining, before the trial, the “background circumstances [of the accused] and contextualizing patterns of offending” were often seen by defence counsel as risking “compromising their client’s presumption of innocence” (p. 103).

In the end, the lawyers recognized that these provisions – however well-meaning they might be – were “ultimately insufficient to address the structural inequalities that Indigenous peoples face in navigating a settler colonial justice system” (p. 103).

Conclusion: Other research has demonstrated the important role that pretrial detention plays in the justice system (see the Criminological Highlights special issue). This study found that the extent that Indigenous identity informed bail decisions “varies widely, [and is] shaped by localized court cultures, systemic constraints and judicial orientation” (p. 105). “These findings point to the limitations of relying on legal principles to address over-incarceration [of Indigenous peoples] in the absence of structural reform – including increased accountability, access to sociolegal resources, and sustained investment in culturally appropriate services” (p. 105).

Reference: MacDiarmid, Laura and Carolyn Yule (2026). Equitable Pretrial Justice for Indigenous Accused? Exploring the Role of Gladue at Bail. Canadian Journal of Law & Society, 41(1), 86-108.

Item 4

Two independent prosecutor-led diversion programs for those charged with low level felony and misdemeanor offences in two locations in the US state of Indiana demonstrate that community programs taking up to a year as a substitute for formal prosecution can reduce reoffending.

All jurisdictions face the challenge of how best to handle minor criminal offences. Some cases are prosecuted criminally, but in most jurisdictions many charges are ultimately dropped rather than being dealt with in criminal court. This paper examines the effect of two such pretrial diversion programs – operated by prosecutors in one moderate sized and one large county in the US state of Indiana.

The two programs varied in their requirements and typically had fees that participants were required to pay (in addition to any restitution related to their offences).  This paper looked at recidivism – defined as a new charge being filed against those being followed in the study. Two groups were compared in each of these two locations: those who were given the opportunity to participate in the prosecutor-led diversion programs who actually enrolled in it (and may or may not have finished it) and those who experienced normal prosecution for their charge(s). It should be noted that those who did not participate in the prosecutor diversion program did not always receive full prosecution. The dismissal rates for those who did not participate in the formal diversion program varied between 42% and 47%. In contrast, 97% of those who were assigned to the prosecution diversion programs resulted in having a dismissal of their charges.

A long list of variables (e.g., race, sex, age, criminal record, nature of the most serious charge) were controlled statistically by way of a complex set of propensity score models. These models were designed to be able to compare statistically equivalent groups of accused people who either enrolled in prosecutor-led diversion programs or who experienced the normal prosecution process.

In both of the two locations that were examined, enrollment in the prosecutor-led diversion program was associated with a reduction in the probability of a new case subsequently being filed against the accused person, though the effect was larger in one county than in the other. This overall finding – a reduction in recidivism associated with participation in a prosecutor-led diversion program – appeared to hold for White, Black, and Latino accused people, though the size of the effect appeared to have varied somewhat across groups. There was some evidence that there were larger effects of the diversion program for justice-involved minorities. 

Conclusion: One of the goals of the prosecutor led diversion programs was to reduce recidivism. The 12-month recidivism for all groups was, in fact, somewhat lower for those in the diversion program compared to the normal prosecution group (many of whom eventually had all charges dropped). The fact that these two programs were somewhat different (only one had active management of the cases) suggests that adult “diversion programs may create opportunities for prosecutors to meet public safety goals while contending with the reality of an ever-changing resource environment” (p. 234).

Reference: Henry, Tri Keah S., Eric Grommon, Carmen Diaz, and Even M. Lowder (2026). Examining the Effect of Indiana Prosecutor-Led Diversion Programs on Recidivism Outcomes: A Case Study of Two Jurisdictions. Criminology & Public Policy, 25, 213-240.

Item 5

The public’s understanding of crime and the criminal justice system are often systematically incorrect. It is not surprising, therefore, that the public’s views about how to respond to crime are inconsistent with evidence about what might be effective.

There is a substantial amount of research reporting systematic errors made by ordinary citizens in their assessments of crime and the operation of the criminal justice system. In addition, there is information that governments respond to these public attitudes (e.g., see Criminological Highlights special issue on public confidence in the criminal justice system and Criminological Highlights 14(4)#2). This report presents findings from a large-scale survey in England & Wales about crime, victims, policing, fear of crime, the operation of the justice system, crime prevention, and crime policy. Some of the main findings are summarized below.

Nature of crime. 

• People (especially supporters of conservative parties in England & Wales) substantially over-estimate changes in the crime rates over the past 10 years of most ordinary crimes (e.g., murder, domestic burglary, vehicle related theft). People were, however, largely correct about an increase in the reported incidence of one crime: shoplifting.
• Most (64%) of those convicted of homicide are white, whereas the public’s estimate is that many fewer of those convicted of homicide (40%) are white. Supporters of the far-right reform parties think that only 30% of homicide offenders are white.
• 12% of English prisoners are foreign nationals, but residents of England estimate that a higher portion (25%) of prisoners are foreign nationals.
• Most (63%) residents think that the number of police has fallen in recent years when, in fact, the number has risen.
• 18% of respondents report that fear of crime has a large effect on the quality of their lives. “Those whose quality of life is highly affected by fear of crime are far more likely to wrongly say that crime has increased since 2010” (p. 34).
• 39% of respondents report that they do not feel safe walking on the street after dark in their areas. Women are more likely to report feeling unsafe than are men (48% vs. 28%).
• 29% of the public think that people from ethnic or religious minority groups tend to receive more lenient sentences than do others for similar crimes. 23% think that ethnic and religious minority group members receive harsher sentences. Conservatives and Reform party supporters are much more likely to think that religious and ethnic people receive more lenient sentences.
• Those who greatly over-estimate the ethnic minority share of people convicted of homicide are more likely (44%) to believe that minorities receive lighter sentences than the UK public as a whole (29%).
• Most people (65%) anticipate that crime will increase in the next 3 years.
• More people have confidence in the police (35%) than the government (20%) to deal with crime.
• People are more likely to oppose (52%) than support (35%) giving harsher sentences to immigrants.  But those who think that foreigners are already a very high proportion of the prison population (more than half) are more likely to support harsher sentences (54%).

Conclusion: It is clear that there is a fair amount of misinformation about crime and the justice system. To some extent, holding mistaken views about crime is correlated with the sources that people use to get information. But more importantly, when people express views about crime and justice policies, it is likely that these are based, in part, on mistaken understandings about crime and justice.

Reference; The Policy Institute, King’s College London (2026). Mistaken Britain: Wrongful Convictions? Misperceptions of Crime Among the UK public. https://www.kcl.ac.uk/policy-institute/assets/wrongful-convictions-misperceptions-of-crime.pd

Item 6

Toronto’s Safe (Drug) Consumption Site program was designed to lower the risk of death from drug use. Crime did not significantly change in the immediate area of the 10 locations in which these safe consumption sites were placed following their opening.

The use of synthetic opioids, including fentanyl, is estimated to have been responsible for thousands of deaths in Canada. Though many countries (notably the US) have declared a “war” on drugs, more recently, in some jurisdictions, harm reduction programs -- including locations where drugs can be taken more safely under professional supervision -- have been established. One concern about such safe consumption sites (SCSs) is that they will increase ordinary crime in the neighbourhoods in which they are located.  The evidence from other studies suggests, however, that SCSs generally do not affect crime.

This study examines the impact of SCSs on crime in Toronto. The SCSs opened over a three-year period (2018-2021). The focus was on the immediate area near the SCS – typically about 18 “street segments” (one side of one city block) housing roughly 400-700 people. Crime was measured in the immediate area of each SCS before and during the operation of the SCSs (from 2014 to 2023 with the SCSs). Many of the SCSs were in low-income areas and, not surprisingly therefore, had crime rates higher than the city average before implementation. “Synthetic controls” were used to evaluate the effect of the SCSs based on pre-implementation measures. In this way changes in crime rates in the SCS locations could be compared to estimates of what occurred elsewhere in the city. Various other characteristics of neighbourhoods (e.g., the number of consumer establishments, the presence of government living assistance and other aid programs) were used to create areas that were matched to the SCS areas. Two aggregate (overall property and violent crime) and 8 specific crime measures were examined.

“Crime was not significantly impacted, in either direction, by the implementation of the SCSs” (p. 549) in the local area containing the SCS. This was the case for all 10 measures of crime. However, there was a fair amount of variability across the ten SCSs. When looking at the individual locations, it was found that each of the 10 crime measures had at least one significant increase, one significant decrease and one non-significant effect, in the various SCSs suggesting that factors other than the presence of the SCS may have had some effect on certain crime rates. Though the overall rate was unaffected by the presence of an SCS in the local area, there was an indication that the exact locations of crimes may have shifted somewhat (violent crime further from the SCS and property crime closer). There was however “no meaningful change in the spatial distribution of “total crime” from the 6 months prior to the 6 months post-commencement of the SCS implementation.”

Conclusion: “The results generally indicate that Safe Consumption Sites do not produce significant criminogenic effects” (p. 567) at the local level. The variability across areas suggests that other factors that may change the areas in the immediate vicinity of the Safe Consumption Sites may be more important than the presence of the site itself. However, “In a general sense, concerns around Safe Consumption Sites and crime may be overstated, at least in terms of changes in crime levels across major crime types at the micro-level surrounding [these sites after implementation]” (p. 571).

Reference: Frederick, Ainsley, Nathan T. Connealy, and Timothy Cubitt (2026). Safe Consumption Sites and Crime: A Staggered Synthetic Control Evaluation of Toronto’s Expanding Safe Consumption Network. Journal of Experimental Criminology, 22, 537-575.

Item 7

Americans of South and Southeast Asian descent are more likely to be stopped and questioned if they have dark skin.

A previous study demonstrated that in police traffic stops of ordinary citizens searches were most likely to take place when White officers stopped Black drivers (searches took place in 8.2% of stops) and were least likely when Black officers stopped White drivers (1.5% of stops) [Criminological Highlights 13(2)#8]. Similarly, Hispanic defendants in Florida who were rated by independent raters as having darker skin were more likely to be convicted and were more likely to be incarcerated than those with lighter skin (Criminological Highlights 21(1)#6). A study carried out in Minnesota showed similar findings. But in addition, accused people who were White and had ‘Afrocentric features’ were significantly more likely to be punished severely than were those who did not have these features (Criminological Highlights 16(3)#1; See also 18(3)#4). This study extends the study of skin colour to another group of Americans – those of Asian descent.

In this study, a sample of Asians – East Asians (Chinese Japanese, Korean, Taiwanese), South East Asians (Vietnamese, Filipino, Cambodian, Thai, Indonesian, and people from 6 other countries), and South Asians (Indian, Pakistani, Bangladeshi and people from 4 other countries) were interviewed. Each was given a scale of skin colour depicting 10 hands of varying darkness and asked to identify the colour that came closest to their own colour. They were also asked whether they had been stopped and questioned by the police in the past 5 years. Various other factors that might be expected to correlate with police stops were statistically controlled. These included whether they were born in the US, their household income, the racial makeup of their neighbourhood, and a number of questions related to their prior involvement with the criminal justice system.

The respondents were divided into four skin-colour categories (Lightest skin, 1 or 2 on this ten-point scale; Light brown – 3; Medium brown – 4; or dark skin -- 5 to 10.) About 18% of the total sample of 2,970 people from all three areas of Asia (East, Southeast, South) had been stopped by the police in the previous 5 years. For Southeast Asian and South Asian respondents – but not for East Asians – those with darker skin were significantly more likely to have been stopped by their local (US) police even when the various social characteristics (e.g., criminal record) were statistically controlled. 

A more detailed analysis suggests that the effect of skin colour of the Asian respondents is best described as an effect of having dark skin (5 to 10 on the 10-point scale). This held separately for the Southeast Asians and the South Asians. For East Asians, however, only 6% described themselves as having “dark skin” (any skin colour between 5 and 10 on this scale) compared to 18% of Southeast Asians and 25% of South Asians. In other words, it is possible that the lack of an effect of skin colour for East Asians relates to the simple fact that there were almost no East Asians with “dark” skin.

Conclusion: “While [Asian-Americans] remain underrepresented in criminal justice [arrest, etc.] statistics, [the] findings suggest that segments of this population may experience racialized policing…. Because they live in a society that punishes Blackness…., those with complexions that signal distance from Whiteness and adjacency to Blackness may also have to endure a disproportionate burden of unwanted police attention” (p. 957).

Reference: Baluran, Darwin A. (2026) Skin color stratification in police contact among Asian-descent people. Social Problems, 73, 941-960.

Item 8

For people who are arrested, the association between the experience of arrest and perceived arrest risk (if the person were to offend again) is stronger for recent arrests than for earlier arrests. This pattern supports the conclusion that the influence of an arrest on perceived risk may fade with time.

Information about what has happened to a person very recently appears to be more likely to affect a person’s expectation of what will happen in the future than temporally distant information. But in addition, the effect of new information appears to erode over time. This paper examines whether the association between arrest experiences and perceived risk of arrest is weaker when the arrest occurred in the more distant past. The authors consider whether this pattern could help explain why the deterrent effects of arrest on offending may be short-lived.

The paper uses data from a longitudinal study that followed a sample of youths (N=1,354) from adolescence into adulthood. The youths’ involvement in the study started with their conviction for a criminal offence in Arizona or Pennsylvania. There were 10 follow-up interviews: the first 6 interviews took place at 6-month intervals and then annually. Each youth was asked how likely it is that they would be caught and arrested if they committed each of seven types of crime. They answered on a scale from 0 (no chance) to 10 (absolutely certain to be caught). On average, across time and crimes, people thought that there was about a 50% likelihood of being caught. The actual number (and timing) of the arrests were taken from official data. A second study used data from Denver, Colorado, in which people who had offended were interviewed four times. The findings of the two studies are very similar.

Various measures related to the deterrent impact of arrests were calculated. First, the researchers calculated the “arrest signal” as the ratio of arrests per crime committed. A respondent was considered to have experienced a “recent arrest” if the arrest occurred in the three months prior to the interview.

People were asked what the likelihood was that they would be picked up by the police if they engaged in violence and theft. Because respondents indicated when they had last been arrested, the researchers were able to determine if they were a recent arrestee – being arrested within three months of the survey – or whether an arrest took place further back in time. 

Respondents who were arrested within 3 months of the recall survey showed a relatively large impact of the arrest on the perception that they would be caught if they committed an offence. The effect of arrest on the perception of being caught is “driven primarily by individuals who were recently arrested who tend to update their perception of risk considerably. [However,] the impact of arrest experiences tends to decay as more time passes” (p. 786).

Conclusion: The data suggest that “individuals update perceptions of arrest risk based on offending experiences” (p. 794). Said differently, following a recent arrest, respondents tend to perceive a higher likelihood of being arrested if they offend again than they do a few months later. “The updating effect associated with experiencing an arrest appears to be relatively short-lived, and individuals tend to “reset” back to their [prior estimate of the likelihood that they will be arrested] not long after being arrested. This study thus provides an explanation for the findings supportive of the short-term specific deterrent effects [of arrest as well as the finding that] the long-term impacts are negligible” (p. 794).

Reference: Thomas, Kyle J., and Greg Pogarsky (2026). Recency and Decay Effects in Sanction Risk Updating. Criminology, 64, 771-798.

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