Criminological Highlights Vol. 22, No. 5 - November 2025

Item 1

Racial inequalities across various domains can be hidden or denied by making race-neutral assumptions about the meaning of the variables that are statistically controlled.

Imagine the situation where a Black person gets a more punitive sentence than a White person for an identical offence, where the White person’s previous criminal record only includes misdemeanor (or summary conviction) convictions and the Black person’s record consists of a similar number of felony (or indictable) convictions for the same offences?  Often such situations are interpreted as suggesting that as soon as “criminal background” is controlled for (felonies vs. misdemeanors in this case), there may be no evidence of racially based inequality. 

The problem, of course, is that this situation may well reflect racialized decisions at earlier stages in the lives of these two people – where the decision to proceed with a more serious form of the offence is the result of a racialized decision. Furthermore, in the future, the Black person, in this example, would have a more serious recent record (based on the sentence which itself is based in part on the earlier racialized decisions). Hence in the future, when “controlling” for the recent convictions, it is almost inevitable that the Black person will be punished more severely. Some examples of such “racial innocence” – whereby discriminatory decisions are hidden behind other “innocent” variables – are obvious. For example, the original US federal sentencing guideline mandated a much harsher sentence for possession of crack cocaine (with a majority of those being sentenced being Black) than for an equal amount of pure cocaine (where most people sentenced were White). By thinking of rules such one as this as being “race neutral… racial inequalities in sentencing stemming from these factors are [seen as being] warranted” (p. 387).

This paper suggests that there are three mechanisms that “enable racial innocence in social science: treating unequal structural conditions impartially, isolating sample choices to reflect narrow stages, and focusing on individual levels of analysis” (p. 388).  The paper uses a dataset from 2012-2015 Miami-Dade County Florida to illustrate the mechanisms of creating “racial innocence”. Race of the person being sentenced was determined by the police categorization on the arrest form. Various personal and legally relevant variables were coded from the data. The main focus of the analyses was on the outcome for the Black and White people being sentenced. 

The effect of race, “taking into account” factors known to be relevant to sentencing (e.g., criminal history, most serious charge, whether the person experienced pretrial detention, etc.) was small, but significant: Blacks were about 3% more likely to be imprisoned than Whites.  But when the legal and case characteristics were assumed to be racialized rather than treating as if they were not related to race, “racial inequality incarceration doubles” (p. 398).  One problem in looking only at the sentencing of those found guilty is that such an approach assumes that being found guilty is not itself a racialized decision. When the researchers looked at the full arrested sample to see who was, in the end, incarcerated, the impact of race on being incarcerated was larger than that of the convicted sample alone. Similarly, when the neighbourhood in which an accused was living was taken into account, the findings showed that “racial inequality in incarceration is more than twice as high in the neighbourhood-level models compared to the individual-level models” (p. 403).

Conclusion: This paper demonstrates that “racial innocence” in the criminal justice system is the result of the manner in which we examine how people are treated by the system. By assuming that factors such as criminal record or pretrial detention are objectively rather than, in part, racially determined, we obscure much of the impact of racialized decisions. It is argued that a race-conscious approach integrated more broadly into criminal justice research would better identify how race-neutral thinking minimizes racial inequalities at all stages of the criminal justice system.

Reference: Omori, Marisa, A. M. Early and L. Torres (2025). A Theoretical and Empirical Critique of Racial Innocence in Sentencing. Law & Society Review, 59, 382-418.

Item 2

The impact of having at least one Black person on a criminal jury far exceeds the direct impact of that juror’s single vote. Black jurors can bring an understanding of the case to the other jurors that might otherwise not be heard.

Restricting criminal jury membership to those who have not had direct experience (or knowledge) of interactions with the police can easily keep juries from understanding legitimate concerns about certain police-citizen interactions. Specifically, restrictions on the inclusion of Black people as jurors can mean that behaviour of an accused person may not be interpreted in a manner that would be easily understood by any Black person who has had interactions with the police.

Using people who volunteered to be on ‘mock juries’, the study examined the impact of the race of the defendant (Black or White) and the race of a key law enforcement-cultivated witness (Black or White) on judgments in a drug conspiracy case. A third key variable was whether the ‘jury’ included one or more Black jurors.

In all, 144 jury groups (consisting of 4 to 7 participants from the Central District of California) were randomly assigned to one of the four conditions (the accused and the prosecutor’s key witness were described as being either White or Black). The case involved a charge of conspiracy to distribute more than 100 grams of heroin.

The mock jury’s deliberations were coded on whether or not there were positive and negative discussions about the law enforcement witnesses. When there was at least one Black juror on the jury, there were significantly more negative discussions about the law enforcement testimony than when there were no Black jurors. This was the case both when the defendant was Black and when the defendant was White. The number of positive discussions about the law enforcement evidence was higher when there were no Black jurors. When there was at least one Black juror in the group, there were fewer favourable discussions about the credibility of the police officers’ testimony and more discussions questioning the credibility of the police officer testimony.

In an attempt to understand how groups moved, in their deliberations, toward acquittal, the deliberations were examined. Jurors sometimes mentioned how they, if they were stopped by the police, would have appeared nervous even if there was no issue of drugs or other criminal matter. They pointed out that, given relationships between the Black community and the police, it was completely normal for a person to be nervous. Hence being nervous did not necessarily signify anything about a criminal offence. Similarly, in their testimony, the police officer mentioned the fact that the accused had his papers (driver’s license, insurance, etc.) ready before being asked demonstrated an unusual level of cooperation which was an indication that the person was attempting to avoid further investigation. A Black juror mentioned that this was what one does: “You’re just being Black. That’s the reason you’re nervous” (p. 434).

Conclusion: In their deliberations, Black jurors appeared to interpret the evidence presented in the trial in light of their own experiences. Black jurors, for example, raised the very real problem of being a Black person who was stopped by the police, including the “disproportionate likelihood of being pulled over, and the risks of being harmed or killed in those encounters” (p. 441). Essentially, as jurors are expected to do, Black jurors interpreted the evidence in terms of their own life experiences. This is, of course, the reason that prosecutors are sometimes motivated to try to avoid having Black jurors on cases such as this one. Doing so, however, deprives the jury of hearing an explanation of an accused person’s behaviour that does not necessarily support a guilty verdict.

Reference: Lynch, Mona & Sofia Laguna (2025). Police Talk in the Jury Room: The Production of Race-Conscious Reasonable Doubt Among Racially Diverse Jury Groups. Law & Society Review, 59, 419-448.

Item 3

The judgement of Americans about police behaviour is based largely on the exact nature of the interactions between police and civilians, and not generally the race of the police officer. “People support punishing civilians who pose threats to order, even though they also view hostile officers as less fair” (p. 355).

Developments in technology have allowed ordinary members of the public to see, and judge, the behaviour of police toward civilians. This paper examines how ordinary members of the public view police behaviour. The focus is on the race of both the police officer and the citizen in the interaction and whether the participants addressed the other person in a polite or hostile fashion.

Vignettes were given to US residents who were volunteers for an online study in 2021 and 2024. Individuals responded to only one vignette.   In the first study, the police officer was described as either Black or White. The citizen was also described as being Black or White. The situation was that the police officer had been dispatched to investigate possible gunshots in a park that was supposed to be closed. It turned out to be fireworks.  The police officer attending the scene initiated discussion in either a polite or hostile fashion and the citizen responded either politely or in a hostile fashion. The outcome of the interaction described in the first two experiments was that the police officer asked the citizen to leave the park or the police officer gave the citizen a ticket for disorderly conduct.

Some of the findings from the three experiments suggest that compared to the situation where both were described as being polite, people preferred the harsher punishments when the civilian was hostile and the police officer was polite. In all experiments, if the civilian was polite and the officer was hostile, the preferred punishment was no different than if both were polite. But “when civilians are hostile [toward the police officer], the public wants them punished more” (p. 345). “No evidence showed that the public wants lesser sanctions merely because an officer is hostile while a civilian is polite” (p. 345). Officer race did not seem to be very important. However, “When the civilian initiates hostility…  preferred punishments are significantly higher when both the officer and civilian are white” (p. 345). 

In the final experiment, the police were described as being dispatched because of a report of a fight in a park. Citizens in one set of vignettes had a weapon (a baseball bat). As in the other experiments, citizens and the police officer were described as being polite or hostile. The threat was seen as being higher when the citizen was hostile and the officer was polite.  Respondents were asked not only about their preferred outcome, but also whether they thought that the outcome that was given (citizen arrested or ticketed and told to leave the park) was fair.  Hostility on the part of the officer did not affect the ratings of the preferred outcome, but it did reduce assessments of fairness.

Conclusion: “Even though the mass public’s punishment preferences and fairness judgements [of interactions between individual police officers and ordinary citizens] are generally not affected by officer’s race, they depend on various civilian and officer behaviours” (p. 354). “When officers and civilians are courteous and civilians are unarmed, the mass public prefers that an interaction end in less punitive ways that nonetheless involve active police involvement. Police-civilian interactions involving hostility and threat… are judged differently…: While hostile officers are evaluated negatively, hostile and threatening… civilians are perceived to be less deserving of lenient treatment” (p. 354-5).

Reference: Vaugh, Paige E. and Gregory A. Huber (2025). Seeing the State in Action: Public Preferences about and Judgments of Common Police-Civilian Interactions. Criminology, 63, 330-381.

Item 4

A new scale for predicting violence after release from prison shows that those assessed as having a high risk of committing violent acts after release were five times more likely to commit a violent offence than those assessed as being low risk of committing violence. Nevertheless, 89% of those assessed as being high risk for committing violence if released did not, in fact, commit a violent act within 12 months of release.

Prisoners and those accused of crimes are often subject to predictions about what would happen if they were to return to the community. This happens most clearly in the case of bail/pretrial release: Release is based largely on what they are expected to do. But it also happens when a prisoner is being considered for conditional release from prison or when consideration is being given to placing conditions on a release back to the community.

Hence it is not surprising that criminal justice agencies often attempt to develop better instruments for predicting behaviour – especially violent behaviour – for use with people who are being returned to the community. This paper examines a risk measure, developed for Dutch prisoners, that can be administered by trained general prison staff (rather than psychologists or psychiatrists).

The total of 1320 former detainees were followed sufficiently long to collect data on reoffending in the 12 months following release. The scale – administered while the person was in prison – is quite elaborate, looking at violence prior to imprisonment and while in prison, four measures of dynamic risk factors and four dynamic protective factors. Clearly this is much more information than is normally available for some people (especially those where detention before trial is being contemplated).

A commonly used index of accuracy used in the literature of predicting future behaviour of those involved in the criminal justice system is the AUC which, psychologists often suggest, is of “medium” accuracy if the score is between 0.66 and 0.71 and “large” if it is above 0.71. In this study, the AUC for violence in the community after release within 12 month was 0.68 (“medium accuracy”) and 0.72 (“high accuracy”) for violence within 6 months. Another way of describing the accuracy of the predictions was that the 6-month reoffending rate was 9 times higher for those with “serious concerns” for reoffending compared to those with “low concerns” (1% vs. 9% reoffending). The ratio for 12-month reoffending was that the risky prisoners were 5 times more likely to commit a violent offence (2.3% vs. 11.4%).

These figures might be interpreted as showing strong support for the use of this predictive instrument except for one set of findings highlighted in the paper: the “false positive” rates. In this case, the proportion of people assessed as having a high likelihood of violent offending after being returned to the community, but who did not engage in violence, was extremely high. Specifically, 92% of the “high risk” people did not commit violent offences within 6 months and 89% of these same “high risk” people did not commit a violent offence within 12 months (Table 4). Simply put, most of those predicted to be dangerous turned out not to be.

Conclusion: The paper demonstrates that apparently very predictive scales can disadvantage an enormous proportion of those it suggests are high risk. In this case, if negative consequences were attached to those who were “high risk” it would mean that approximately 90% of those who experienced these negative consequences (e.g., punitive conditions or delayed release) would not have committed violent offences if they had simply been released. Clearly, any scale described as making “accurate” predictions, such as this one, needs to be examined carefully (as was done and reported in this study).

Reference: Smeekens, Marjam V., M De Vries Robbe, A. Popma, and M.M. Kempes (2025). The Prospective Prediction of Community Violence after Release from Prison with the Risk Screener Violence (RS-V). The Prison Journal, 105, 350-374.

Item 5

When ordinary citizens call the police, an in-person meeting of the citizen with a police officer is often not required. A study comparing in-person to virtual meetings with police demonstrated that virtual meetings are often not only more efficient but are seen in a more favourable light by ordinary citizens.

Many calls to police require some form of follow-up by police officers, though if the call does not require urgent police action, alternatives to a formal in-person visit by a police officer may be more efficient. This paper reports a study in which a subset of citizen calls to the police was randomly assigned to receive either an in-person visit by a police officer or a virtual visit. 

In one of the local policing areas in Dorset, U.K., the police received, in 2022, 69,816 calls that required police contact but did not require an immediate response. On average the police-citizen contact took place an average of 67 hours after the call was received. Instead of presuming an in-person meeting in response to all calls, this study compared in-person and virtual meetings. If the call handler determined that a call from a citizen was not an emergency but did require further police contact, the call handler asked the citizen whether the caller would be willing to have a virtual follow-up contact with the police. If the caller was willing to have a virtual rather than an in-person follow-up, the case was then randomly assigned to receive one or the other type of follow-up. The virtual follow-up involved a two-way video communication system that the citizen could connect to, at an agreed time, via a text message or an email.

Not surprisingly, the total time (including travel time) spent by the police was considerably more in the “in-person” condition than in the virtual meeting condition. Witness statements were much more likely to be taken during the preliminary investigation in the virtual condition. Victims were more supportive of the treatment they got from the police in the virtual condition perhaps because the analysis of the victim’s needs analysis was completed more quickly. It was estimated that the total police cost of the cases randomly assigned to the in-person treatment was approximately three times that of the cases handled virtually.

In interviews carried out after the cases were completed, citizens’ overall opinion of the police after the incident was significantly higher in the virtual condition than if they received an in-person meeting. Those who interacted with the police virtually were more satisfied with the outcome of the case. They also had more confidence and trust in the police and were more likely to report that the police had acted in a procedurally fair manner.  The favourable response of the public may, in part, have reflected the fact that in situations in which an immediate police response is not necessary, “victims are [often] left waiting for days or even weeks without an investigation commencing” (p. 324). In this case, those randomly assigned to receive a virtual contact with the police not only had their first formal contact sooner but were much more satisfied with the process.

Conclusion: This study demonstrates that in situations in which it is appropriate, a virtual, rather than in-person meeting between citizens and police will not only save money but will be much more likely to satisfy members of the public and create favourable views of the police. But in addition, the study demonstrates that in certain circumstances police services, if they are interested in improving their services and their reputations, can carry out random-assignment studies to test the efficacy of procedural changes in the manner in which they interact with the community.

Reference: Gates, Stewart, Barak Ariel, and Noy Assaraf (2025). Responding to Nonemergency Calls for Service via Video: A Randomized Control Trial. Criminology & Public Policy, 24, 309-331.

Item 6

Firearms prohibition orders are often imposed on those charged with domestic violence offences. This study demonstrates that, even when they are mandatory, these orders are sometimes not imposed. White defendants in domestic violence cases are less likely than Black and Latino defendants to have firearm relinquishment orders imposed on them.

Because of concern about repeat offending from those charged with a domestic violence offence, several US states have instituted civil laws requiring those charged with a domestic violence offence to relinquish any firearms they possess.

The treatment of firearms in the US has, historically, been a racialized legal matter. In the early 1800s, some state laws prohibited the possession or carrying of firearms by Black citizens. This study examines whether Domestic Violence Protection Orders, requiring those charged with a domestic violence offence who have access to firearms to turn them over to legal authorities, are enforced equally for White vs. other racialized groups. 

Court records were used to identify accused people charged with a domestic violence offence and who, when asked, indicated they had access to a firearm. White accused people were less likely to be ordered to relinquish their firearms. One way to see the size of the effect is to control for other differences in the groups and look, for example, at the probability of a relinquishment order being ordered in the year 2016 (when the order was mandatory) for a 38-year-old male respondent. The probability of being ordered to relinquish firearms was lower for White accused people (80% received such orders) than it was for Indigenous (84%), Black (83%) or Latino (84%) accused people.

In circumstances where the imposition of the order to turn in one’s firearms was discretionary, the difference across racialized groups was larger. For White accused people, 28% were given such orders, compared to 34% for Indigenous, 38% for Black and 45% for Latino accused people.

The race of the person in question was determined by how the accused person was perceived by others (those responsible for the court records). Given that the focus of the study was on disparity of treatment imposed by others (i.e., the court), this measure would appear to be appropriate in determining whether differential treatment was imposed on different racialized groups. 

The data show that White respondents had the highest reported access to firearms, yet they were less likely to be ordered to relinquish those firearms.

Conclusion: The data provide evidence of two effects. Statutory requirements can increase dramatically the imposition of the requirement that accused people should turn in their firearms. And statutory requirements can reduce – but not eliminate – the disparity across racialized groups. However, it is clear that attempts to remove access to firearms for Indigenous, Black, and Latino suspects were more pronounced than for White accused.

Reference: Kafka, Julie M and 7 others (2025). Disparities in Court Orders to Relinquish Firearms in Civil Domestic Violence Protection Orders. Criminology & Public Policy, 24, 405-427. 

Item 7

Having a criminal record dramatically decreases the likelihood that a person will be offered a job that they apply for. This appears to be the case both for those who have been described, officially, as having a criminal record and those who appear to have one as a result of a “Google search.”

Experimental research published in the past 20 years has shown that people with criminal records are disadvantaged in getting jobs, housing or admission to universities (Criminological Highlights 6(3)#2, 21(2)#6), 17(2)#6, 18(4)#2, 15(1)#7, 21(5)#6). There is some public support for reducing the impact of a criminal record and there are techniques to overcome the effect of records (19(3)#1), 22(4)#6, 18(3)#6, 20(3)#4). However, overcoming the impact of a record is not a simple task (21(2)#7, 16(3)#2, 19(4)#5).

This paper makes an additional important point: unofficial public information that a person has a criminal record (e.g., obtained via a Google search) can also have a negative impact on those with criminal records.

In this study, a survey was carried out of people who, at one point in their lives, had jobs involving the hiring of employees.  As participants in an online survey carried out in 2021, they were asked to assess an applicant’s materials for a position of a front desk clerk at a hotel in Cleveland, Ohio. This “job applicant” either had an official criminal record for two minor offences (trespassing 7 years earlier, and petty theft 5 years earlier) or respondents were told that there was no official criminal record for the applicant.  They were also given results of a Google search that showed either that a Google search had found that the applicant had apparently been arrested for these offences, or that the search had found nothing relevant. Respondents were described as being either Black or White.

For both Black and White applicants, those with official criminal records were less likely to get favourable assessments than were those without official criminal records. The effect of the unofficial “Google” record of offending by the applicant was significant – those with an unofficial record were rated less favourably (regardless of their official record) but the effect was smaller than the effect of the official record. The existence of a record (“officially” demonstrated or reported as a result of a Google search) reduced the willingness of people to hire the job applicant, though the effect of the evidence of an official record was larger for both Black and White applicants. The effect of a criminal record on willingness to hire the applicant was larger for White applicants than it was for Black applicants.

Conclusion: A criminal record is highly stigmatic independent of the applicant’s race. The effects of a Google hit suggesting a history of offending reduces the job applicant’s chances of being hired suggesting that simple official suppression of a history of committing minor offences such as the ones used in this study may not be sufficient if evidence of previous offending is available to anyone with a computer.  An encouraging finding, however, is that when one looks at the US evidence in studies carried out over the past 20 years, there is some indication that the impact of a criminal record on job prospects may be lower than it was at the turn of the 21^st century.

Reference: Lageson, Sarah and Robert Apel (2025). The Mark or Trace of a Criminal Record: A Survey Experiment of Race and Criminal Record Signaling. Criminology,63, 382-410.

Item 8

Job applicants’ criminal records can be communicated to potential employers when their job applications indicate that their vocational credentials were obtained as part of a program for prisoners. Having such formal qualifications does help former prisoners get relevant jobs. But the stigma of having a criminal record does not disappear.

There is substantial evidence that having a criminal record substantially decreases the likelihood that a person will be offered a job (see article 7 in this issue). This paper examines the possibility that the stigma of a criminal record might be overcome if the job applicant completes a relevant vocational training program while in prison.

The possibility exists that even though a criminal record may normally signal negative characteristics of the job applicant to a potential employer, the successful completion of a job skills program in prison may erase the negative impact by demonstrating that the prisoner has changed. This study examines the impact of successfully completing a job training program in “heating, ventilation, and air conditioning” (HVAC) – a program often offered in prison in part because of the demand for people with these skills.

The study used job postings from 1502 employers hiring HVAC workers in 5 US states. The employers did not require a specific vocational credential. All applicants were described as having 4 years of work experience and high school skills. Applicants were described either as having a 4-year prison record involving drug offences or there was no mention of any involvement with prisons.  Applicants also were described as having successfully passed an HVAC program (in prison or in the community, depending on the condition) or there was no mention of this credential.  Finally, applicants were described (by using names associated with race) as being either Black or White. The key outcome variable was whether or not the employer responded positively (e.g., requesting more information or an interview with the applicant).

In the analyses, various controls were included. For example, because the states differed in the manner in which they controlled the use of criminal records, the state was controlled as was the month in which the application for employment was filed and whether the employer required a professional reference. 

The analyses showed three main effects. Those with criminal records were less likely than those without criminal records to receive positive responses from employers. Black applicants were less likely to receive positive responses than White applicants. And those with HVAC credentials were more likely to receive positive responses than those without such credentials. The effect of HVAC certification and prison records were very similar for both Black and White applicants. However, for each of the groups (HVAC or not; prison record or not) equivalent Black applicants were less likely to receive a favourable outcome from a potential employer than were White applicants.

Conclusion: The results showed that while “vocational credentials can partially mitigate employment barriers faced by returning citizens, they are insufficient to fully overcome the additive effects of prison record stigma and racial discrimination” (p. 427). “Racial discrimination effects were comparable in magnitude to prison record effects, creating compounded disadvantages for Black returning citizens…” (p. 428). Vocational credentials, then, can help both Black and White prisoners get jobs after they are released. However, at least in the case of HVAC credentials, there is no evidence that vocational credentials put Blacks and those with criminal records in position comparable to White applicants or those who have not experienced imprisonment.

Reference: Lindsay, Sadé L. (2025). Mixed Signals from Prison? Postsecondary Vocational Credentials, Race, and Postrelease Employment. Criminology, 63, 411-436.

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