View this issue as a PDF: 
CrimHighlightsV16N5.pdf
This issue of Criminological Highlights addresses the following questions:
- Why is it in the public interest to allow people to free themselves from their criminal records?
- What kinds of events are likely to cause an increase in police use of force in dealing with ordinary citizens?
- Do highly informed citizens think that sentences are too lenient?
- What can be done to create more smoothly running prisons?
- What is the first step that cities should take to prepare for events that might involve citizen protests?
- Do transfers to adult court hurt youths’ life chances?
- How accurate are predictions of future intimate partner violence?
- Can crime be stopped by increasing the likelihood of apprehension?
Allowing people with criminal records to work in jobs that are presumptively restricted to those without criminal records reduces the likelihood that they will re-offend.
There is a substantial amount of information suggesting that people with criminal records have a difficult time getting employment and housing. Obtaining employment is one of the reasons that people attempt to have their criminal records suppressed (Criminological Highlights 6(3)#2, 15(3)#5, 15(2)#6, 16(3)#2). This paper examines the question of whether allowing people to avoid the negative impact of a criminal record on obtaining employment reduces subsequent re-offending.
Specifically, the paper examines the impact of waiving a normal prohibition against hiring people with criminal records (i.e., arrests or convictions). It focuses on a specific type of employment in which many job applicants have criminal records: Low wage jobs in the health care industry for which people do not need a specific license (such as those required of doctors or nurses), but do have contact with patients. These were generally unskilled jobs in facilities such as nursing homes, assisted living residences, etc. 71% of the applicants were women; 50% were Black.
The criminal background check for those wishing to be employed in these jobs in New York State involved a two stage process. A person could apply for the position, and then, if they had no record (only 48% of all applicants), they would be cleared for the job. Alternatively, a person with a record would normally be initially and tentatively denied the position, but would be invited to provide evidence of rehabilitation to reverse the proposed denial. “About 59% of those with criminal records did submit evidence of rehabilitation. 52% of those who submitted evidence of rehabilitation were cleared for employment. Those who had criminal records but did not submit evidence of rehabilitation were automatically denied the possibility of employment.
Information on criminal justice involvement of all job applicants was collected for the 3 years after the application was submitted. Only about 9% of the job applicants with criminal records reoffended, though this rate was, not surprisingly, considerably higher for men than for women (approximately 11.2% for men and 7.1% for women). Older job applicants were less likely to reoffend.
There was a 4% reduction in reoffending over the 3-year period for those cleared for work. In the context of a reoffending rate of around 10%, this reduction is obviously quite substantial. Given their higher base rate of offending, it is not surprising that the beneficial effect of being cleared for work was larger for men than for women. There was essentially no impact on being cleared for work for older women, a finding that also isn’t surprising, given that they had low reoffending rates overall, whether cleared for employment or not.
Most, but not all, of those cleared for work did, in fact, get jobs. It appears that the large reduction in reoffending by men who were cleared for employment “can be explained by increased employment opportunities in the health-care industry that would otherwise be unavailable” (p. 196).
Conclusion: Enabling those with criminal records to obtain employment that would otherwise be unavailable to them appears to reduce subsequent offending. The job applicants in this study were, on average, about 39 years old. Previous research (Criminological Highlights 4(3)#6) has suggested that employment for those over age 27 is particularly likely to reduce reoffending. It would appear that there are advantages to both those who have offended and society more generally to reducing the barriers to employment for those once arrested or convicted.
Reference: Denver, Megan, Garima Siwach, and Shawn D. Bushway (2017). A New Look at the Employment and Recidivism Relationship Through the Lens of a Criminal Background Check. Criminology, 55(1), 174-204.
The use of force by New York City police officers against Black suspects in stop-and-frisk operations increased immediately after two fatal shootings of police officers by Black suspects.
Explanations that are offered for police use of force against civilians usually relate to characteristics (including race) or behaviour of the civilian or characteristics of the police officer involved.
In contrast, this study suggests that “racial bias in policing and discrimination more broadly is not static but fluctuates, partly driven by significant events that provoke intergroup conflict and foreground racial stereotypes” (p. 380). This paper looks at police use of ‘ordinary’ force against citizens during the two week period before and after two New York City (NYC) police officers were killed by Black suspects. In addition, the use of force was measured before and after two other incidents in which NYC police officers were killed by Hispanic and White suspects. During the period that the data were examined for this study (2006-2012) Blacks were estimated to be about 25% of the population of NYC, but were the subject of about 54% of the recorded stop-and-frisk operations.
Two police officers were shot and killed by Black suspects in July 2007 and in December 2011. In March 2007 and in March 2011, police officers were shot by a white and a Hispanic suspect, respectively. “Use of force” by police officers during stop-and-frisk operations was estimated from the data on the NYC police stop-and-frisk report form. Various forms of physical force were listed on the form. This study coded any use of force (“Hands on Suspect” or more) as use of force. For Blacks, 22% of the stops involved this level of force; for Hispanics it was 24%; for Whites it was 16%.
Comparisons were made on police use of force in stops in roughly the same location and at approximately the same time of day three days before and three days after each of the killings of police officers. The data show that there was “increased use of physical force by police officers against Blacks substantially in the three days after the shooting” (p. 395). Use of force by the police against Black NYC residents subject to stop-and-frisk increased from 21.9% before the killing of the white police officer to 25.4% after the 2011 event. For the 2007 event, police use of force against Blacks increased from 22% (before) to 25% (after).
The use of force against White and Hispanic suspects did not change from the period before to the period after the killing of the White police officer by the Black suspect. Furthermore, “the two [killings of White police officers] involving a Hispanic and White suspect did not increase the use of force against any of the groups” (p. 397).
Detailed daily estimates on the use of force demonstrated an immediate increase the day after the killing of a police officer by a Black suspect, and then a fairly rapid decline such that within about 2 weeks, the rate of use of force had returned to its previous levels.
Conclusion: The findings suggest that racial bias in policing is not static. Its rate is “partly driven by significant events that provoke intergroup conflict and foreground racial stereotypes…. Extreme violence against police officers can lead to periods of substantially increased racial disparities in the use of police force” (p. 410). The findings suggest that “local events create intergroup conflict, foreground stereotypes, and trigger discriminatory responses. From this perspective, discriminatory behaviour arises not only from static conditions but also from temporal sequences of events and responses” (p. 410).
Reference: Legewie, Joscha (2016). Racial Profiling and Use of Force in Police Stops: How Local Events Trigger Periods of Increased Discrimination. American Journal of Sociology, 122(2), 379-424.
An Australian study demonstrates that ordinary citizens who are intimately aware of the details of a specific criminal case are much more likely to agree with the sentence handed down by the court than public opinion surveys would suggest.
Research in many countries suggests that ordinary citizens believe that sentences handed down in criminal courts are too lenient. The irony of this view, of course, is that few people have enough detailed information about criminal cases to evaluate whether the ‘sentence fits the crime.’ They typically have little information about the details of the crime or about the offender.
Previous research, including one prior Australian study (Criminological Highlights, 11(6)#2, 14(1)#6), has found that jurors who have found an accused guilty are much more content with the sentence handed down in the case they heard than public opinion surveys would suggest. Indeed, in many cases, jurors’ preferred sentences were more lenient than that imposed by the judge. Jurors, as representatives of the public at large, then, do not differ much from judges in their views of appropriate sentences.
This study, replicating the earlier Australian study, was carried out in the state of Victoria. It used data from 987 jurors who served on juries in 124 trials. Most of the principal offences for which the accused was found guilty were sex offences (39%) or violent offences (32%). Jurors were invited by the trial judge to participate in the study in trials where a guilty verdict was returned but before the offender had been sentenced. Those who voluntarily participated were then asked if they would be willing to participate in a second survey (after the judge had sentenced the offender).
For 62% of the jurors, the preferred sentence – suggested by the jurors before the offender was actually sentenced – was more lenient than the sentence that was subsequently imposed by the judge. In 2% it was the same and for 36% of the jurors their preferred sentence was more severe than the sentence imposed by the judge. Not surprisingly, the jurors who filled out the follow-up questionnaire thought that the judge’s sentence was either very appropriate (55%) or fairly appropriate (32%). In cases where offenders were sentenced to prison, jurors who thought a prison sentence was appropriate recommended a prison sentence that was, on average, 12 months shorter than that imposed by the judge.
The exception to this general finding was that for sex crimes, jurors tended to equally split between those who wanted a more lenient and those who wanted a harsher sentence than that imposed by the judge. In particular, jurors tended to be harsher than judges for sex offences involving victims under 12 years old. This effect was similar to that found in the earlier Australian study.
When asked about sentences generally, however, the results were quite different. These jurors thought that sentences in their state were, on average, too lenient for all categories of offences.
Conclusion: The method used in this study – asking jurors about their views of the sentence they thought should be imposed in the case they heard – “confirms earlier findings suggesting that top-of-the-head surveys showing dissatisfaction about judicial leniency cannot be taken at face value and reveal that the community is not as uniformly punitive as polls suggest” (p. 198). Jurors, it seems, are quite happy to give their views of what should happen to the person they just found guilty and a substantial portion of them are willing to give their views of the judge’s sentence when it was handed down. They generally are fairly content with actual sentences in the case they heard. Interestingly, however, jurors do not appear to generalize from their own experiences to sentences across the board: they still believed that most sentences are too lenient.
Reference: Warner, Kate, Julia Davis, Caroline Spiranovic, Helen Cockburn, and Arie Freiberg (2017). Measuring Jurors’ Views on Sentencing: Results from the Second Australian Jury Sentencing Study. Punishment & Society, 19(2), 180-202.
What determines whether prisoners see the resolution of a dispute as legitimate? It is not the outcome of the dispute. Rather, it is whether the correctional officer is seen as acting in a procedurally fair manner.
It is believed that “when individuals believe authorities are legitimate, they are more likely to accept and comply with the decisions of those authorities regardless of their self-interests…. If individuals view authorities as illegitimate, then they may be more likely to become defiant or disrespectful toward authorities, which could be linked to noncompliance” (p. 683). Said differently, the legitimacy of those in power in institutions like prisons is important for the peaceful operation of the institution.
This paper examines the determinants of 5,616 prisoners’ views of correctional officers’ legitimacy in 46 correctional facilities in two US states. The prisoners had all served at least 6 months in the facility. A subset of 1,856 of these prisoners reported they had been confronted by correctional officers who alleged they had violated a prison rule. These prisoners were asked questions about two aspects of this dispute: (1) The favourability of the outcome. This was measured with four questions related to such factors as whether the outcome of the dispute was better, the same as, or worse than the prisoner expected. Prisoners were also asked (2) whether the correctional staff showed ‘procedural justice’ in dealing with the dispute by such actions as being polite and honest, showing concern for the prisoner’s rights, using fair investigatory procedures, and giving the prisoners a chance to give their side of the story. The focus of the study was the extent to which these two measures related to the prisoners’ views of correctional officers, generally, as fair, doing a good job, and treating prisoners equally.
Looking first at all inmates (those who had and had not been confronted for a rule violation), the study found that those confronted by correctional staff for a rule violation were, not surprisingly, less likely to see correctional officers as generally acting legitimately. This finding held even after 15 other variables (including various characteristics of the prisoner, such as race or whether they were gang members) were controlled for.
More important was the finding – looking only at those prisoners who had been confronted for a rule violation – that those who thought their own dispute was resolved favourably were no more likely to view correctional officers as generally acting legitimately than those who thought the correctional officer resolved the dispute in a manner that was not in their favour.
What was important for prisoners’ views of correctional officer legitimacy was whether they felt that the dispute had been handled in a legitimate, procedurally fair manner. In other words, “an instrumental perspective on prison discipline that [suggests that prisoners] who receive more favourable outcomes will be more likely to view correctional officers as legitimate” (p. 699) was not supported by this study. Instead, it was found that prisoners, like everyone else, thought it important they were treated fairly and respectfully; this fair, respectful treatment led to the view that officers were doing a good job.
Conclusion: This study found that prisoners who “perceive the outcomes and treatment they received during their encounters with correctional officers as more fair will, in turn, hold stronger beliefs regarding the legitimacy of those officers” (p. 699). This is important, in part, because “correctional officer legitimacy has… been associated with the stability and predictability of prison environments.
Such environments are more likely to facilitate inmate well-being” (p. 698). In other words, procedural fairness is important for “improving the safety and well-being of inmates and staff, and also for facilitating behavioural change among inmates” (p. 698).
Reference: Steiner, Benjamin and John Wooldredge (2015). Examining the Sources of Correctional Officer Legitimacy. Journal of Criminal Law & Criminology, 105(3), 679-704.
Major events in large Canadian cities may be less likely to result in disorder and arrests of large numbers of citizens if coherent and comprehensive legal frameworks are developed before the events take place.
A number of reviews have been carried out of disruptions at large events such as the G20 meeting in Toronto in 2010. These events are sometimes seen as ‘policing failures.’ This paper explores potentially controversy-attracting events and suggests, instead, that attention should be focused, initially, on the complex legal frameworks in which they take place.
This paper takes what is known about the policing problems uncovered after the disruptions at the Toronto G20 meetings and explores what could be learned from a very different event: the winter Olympics in Vancouver in the same year. One symptom of the problem in Toronto is that 1,105 people were arrested in connection with protests at the G20; however, only 117 were ever found guilty of anything.
Among other challenges facing those in charge of making arrangements for the G20 was the fact that Toronto had only 4 months to plan and implement procedures. Four concerns had to be simultaneously addressed: security for the event itself, management of vehicle and pedestrian traffic for those in Toronto, facilitation of collective expression, and law enforcement (e.g., apprehension of violent or destructive people, whether politically motivated or otherwise).
The problem with relying on common law power of the police to “preserve the peace” is that it does not define the limits of what can be done, and in specific circumstances the common law does not define, even geographically, where ‘special’ powers might appropriately be exercised. Thus, for example, it provides no help on defining where the public can be excluded from normally public spaces. Hence when three different degrees of restrictions were placed on parts of Toronto, there was no clear authority for doing this.
The Toronto G20 also suffered from the fact that there was “inter-agency confusion regarding police powers” (p 197). The Toronto Police Service, for example, was uncertain of its authority to construct security perimeters. In contrast, the Ontario Provincial Police believed their exclusion zones around the G8 meetings in Huntsville, Ontario (which took place immediately before the G20) were justified by the common law. A little-used provincial law was resurrected to accomplish this without publicity, making it impossible for members of the public to obey a regulation that almost nobody knew about. More generally, lines of authority were never clear among the various quite independent federal, provincial, and municipal police authorities. Describing the inter-agency problems as a ‘failure of communication’ is an understatement, since it was not clear what was to be communicated or by whom.
In contrast, in Vancouver, planning for the Olympics was comprehensive and carefully thought out. Public debate, long before the event, was sparked by proposals concerning the restrictions on public access (for protest/demonstration purposes). Legal gaps were filled with a provincial law and a municipal bylaw (which was subsequently amended to address certain concerns); jurisdictional issues relating to the two police services involved were worked out far in advance. Debate largely occurred before, rather than during the events in large part because all issues were in the open in Vancouver. This was not the case in Toronto.
Conclusion: If the legal structure governing behaviour in large public events is properly structured, citizens and policing authorities will know how the event will be managed. In that way, management of these events can be structured so as to minimize the likelihood that police-citizen relationships are destabilized and individual rights threatened. Prior to the event, then, the authority of the various police services needs to be clarified, and concerns about balancing of interests need to be addressed explicitly.
Reference: Pue, W. Wesley, Robert Diab, and Grace Jackson (2015). The Policing of Major Events in Canada: Lessons from Toronto’s G20 and Vancouver’s Olympics. Windsor Yearbook on Access to Justice, 32, 181-211.
Youths charged with offences in late adolescence and then transferred to adult court are likely to earn less money when they reach their early 20s in large part because they have spent considerably more time in custody than youths who remained in youth court.
There is a substantial amount of information supporting the conclusion that transfers of youths accused of committing crimes to adult court do not have specific or general deterrent impacts (Criminological Highlights, 1(3)#2, 1(5)#5, 3(5)#5, 6(6)#2, 8(4)#7). More generally, transfers do not reduce subsequent offending. These two papers examine whether transferring youths to adult court has an impact on educational attainment and income.
Both papers use data from the “Pathways to Desistance” project that followed youths for seven years after they were found guilty of serious offences when they were 14-17 years old. The project was carried out in two locations: Philadelphia and Phoenix. The study by S (see references) looked at males and females from both cities. The A&L study looked only at males from Phoenix. Each used ‘propensity score matching’ in an effort to equate youths who were transferred or to those dealt with in juvenile court. The exact methods of matching varied somewhat, as did the statistical analyses. Given the similarity of the results across studies, these differences indicate that the results are not a result of a particular operationalization of the constructs.
Both papers examine the impact of transfers on the educational prospects and on employment or earnings of the youths. The S study used a composite measure for employment (including hours worked and income earned) during the first four years after the decision on where the case would be heard or the full seven years. The A&L study, in contrast, looked at average hourly wage and average hours worked during years 6 and 7 (at the end of the follow-up period only).
Taken together, the two studies suggest that the overall educational attainment of those transferred and those allowed to remain in juvenile court were similar, though the transferred youths were more likely to receive a high school equivalency rather than a high school diploma. On income, the A&L study suggests that, in years 6 and 7, the youths who were transferred made less money. This, however, is apparently because the transferred sample spent, on average, a higher proportion of those two years out of the community (i.e., in custody) compared to the youths adjudicated in youth court. Indeed, the S study, which looked at the 7 year follow-up period as a whole, found that youths who were transferred did, in fact, fare worse on employment outcomes. However, when time in the community was adequately controlled for, there was no overall effect of the court in which the youths were adjudicated. This was true for both the 4-year and the 7-year follow-up periods.
Conclusion: Taken together, these studies suggest that there probably is not an overall impact of transfer to adult court on the level of educational attainment, but transfer clearly does impede the ability of young adults to earn a living, in large part because youths transferred to adult court spend more time incarcerated and less time in the community. One limitation of these studies is that they followed the youths only until they were in their early 20s. Nevertheless, given the known lack of deterrent value of juvenile transfers and the harmful impacts of imprisonment (on youths and adults), the two studies raise serious questions about the value of juvenile transfers.
References: Sharlein, Jeffrey (2016). Beyond Recidivism: Investigating Comparative Educational and Employment Outcomes for Adolescents in the Juvenile and Criminal Justice Systems. Crime & Delinquency (Advanced availability, 2017). Augustyn, Megan Bears and Thomas A. Loughran (2017) Juvenile Waiver as a Mechanism of Social Stratification: A Focus on Human Capital. Criminology, 55(2), 405-437.
A carefully constructed assessment tool for predicting the likelihood of repeat spousal violence is more accurate than police investigators’ predictions. Nevertheless, for those men predicted by this assessment tool to have a high likelihood of future violence, the prediction was wrong in about 80% of the cases: most men did not reoffend.
In responding to cases of intimate partner violence (IPV), police often decide on what actions to take based on their predictions about what the perpetrator might do in the future. Hence, there is appetite for a measure that quickly and easily can help in making a decision as to whether to warn, charge, or arrest and remove someone apparently involved in IPV.
Previous research (e.g., Criminological Highlights 3(2)#7) has suggested that predictions of future IPV are often inaccurate. This paper investigates the accuracy of a prediction measure developed in cooperation with the police in Israel. Among other goals, the study investigated the relative accuracy of police investigators’ predictions of future violent behaviour on the part of those apprehended for IPV. Two general outcome variables were used: any general complaint to the police about the IPV suspect and complaints involving physical violence.
In order to be useful for police, a predictive instrument has to be relatively easy and fast to administer. But more importantly, it has to be based on information that is readily available in most cases. In this case, the final version of the prediction instrument consisted of 45 items related to the offence and the offender. All were to be answered as “yes” or “no”. The ‘critical’ items (e.g., “Are there signs of severe bodily injuries on the complainant?” or “Was she raped?”) were given much higher weights in the final score than the ‘moderate’ items (e.g., “Did the perpetrator pressure the complainant to cancel her complaint?”). IPV “experts” divided the total scores on the prediction instrument into three risk categories.
Police officers responding to 1,133 IPV cases then gathered the information necessary to calculate a prediction score. In addition, police investigators themselves made subjective predictions concerning the future behaviour of the suspect. They were not given the final prediction score, but obviously were very aware of the data that went into calculating it since they gathered it.
For a period averaging 26 months, police records were monitored for repeat offending of the 1,133 suspects in the study. Two types of offending were noted: General complaints from the victim (e.g., harassment, threats, damage to property, violation of court protection orders, etc.) and physical violence or murder threats.
The police officers’ assessment was that about 20% of the IPV suspects were high risk for reoffending. However, their subjective risk assessments did not predict future behaviour on the part of the suspect. About 17.5% of the suspects were, in fact, the subjects of general complaints and about 13.6% were the subjects of complaints about violence. But the investigator’s assessment had no value in predicting future behaviour. The statistical scale did somewhat better. 6.2% of those rated as ‘low risk’ on the scale were re-involved with the police for physical violence compared to 16% of those whose scale value put them at ‘high risk’. The findings for ‘general complaints’ were similar.
Conclusion: Though this assessment tool showed statistically significant predictive value of IPV, one needs to look carefully at the results to see the limitations. 748 people were assessed to be ‘high risk’ by the scale, but only 120 of them in fact were brought to police attention for physical violence and only 153 for ‘general complaints.’ The lesson is clear, and not unusual: Even though a measure predicts future behaviour to a ‘statistically significant’ degree, that doesn’t mean it will do a very good job. In this case, the vast majority of those who were predicted to be ‘high risk’ for IPV showed no subsequent abusive behaviour toward the original complainant.
Reference: Dayan, Kobi, Shaul Fox, and Michal Morag (2013). Validation of Spouse Violence Risk Assessment Inventory for Police Purposes. Journal of Family Violence, 28, 811-821.
One “dirty little secret of deterrence research” (p. 86) is that people do not estimate the likelihood of apprehension the way “rational actor” or other theories predict they should. Sanction risk perceptions, instead, seem to be both imprecise and, in some instances, “produce logical inconsistencies fundamental enough to violate the laws of probability” (p. 103).
Though the likelihood of offending may be related, to some extent, to the perceived certainty of punishment, this assertion ignores the determinants of the perception of apprehension. Simply put, people may not be very good at processing information that is relevant to assessing their likelihood of being apprehended for an offence. Previous research has demonstrated that perceived likelihood of apprehension and actual likelihood of apprehension are quite independent of one another.
In this study, a series of experiments was carried out with online volunteers to see how they estimate probabilities of apprehension. For example, objectively, the more detailed a hypothetical situation is, the less likely it is to occur because its occurrence requires all of the details to occur. In one study people were randomly assigned to three conditions, two of which were: (a) “Please think about drivers in your city/town. If a typical driver decided to drive drunk on a Saturday night, what is the PERCENT CHANCE that he or she would be ARRESTED by the POLICE?” (b) “Please think about drivers in your city/town. If a typical driver decided to drive drunk on a Saturday night, what is the PERCENT CHANCE that he or she would run a STOP SIGN, cause a CAR ACCIDENT, and be ARRESTED by the POLICE?” The second is, objectively, less likely because all three events must occur rather than just the one. In fact, people’s estimate were quite similar (32% and 29%, respectively.
In another study, testing the “anchoring” effect of information, respondents were given one of two versions (using either the higher or lower numbers) of the following scenario: “Suppose that you drive to a new restaurant. When you go to leave the parking lot, you accidentally hit another car, badly damaging its back end. There are a few people in the parking lot, but none are close by. If you decided to just DRIVE AWAY without taking any other action, do you think your PERCENT CHANCE of getting CAUGHT by the POLICE would be higher than [19%, 79%], or lower than [19%, 79%]?” (Half of the participants received “19%” the other half received “79%”). They were then asked “Now thinking about the last question, what specifically is the PERCENT CHANCE that you would be CAUGHT by the POLICE if you drove away?” Those whose initial anchor was high (79%) estimated that their precise likelihood was 49%; those whose initial anchor was low (19%) estimated that their likelihood was only 32%. It would seem that their responses were ‘anchored’ by the information in the question.
Conclusions: The results of a number of experiments (only two of which are described here) in which people made probabilistic estimates of the likelihood of certain events occurring or of their own behaviour suggest that people’s estimates of what they would do, or their likelihood of apprehension for certain offences, does not necessarily follow simple rules concerning probabilities. These studies suggest that objective probability may be less important than deterrence theorists would suggest. Attempts, therefore, to reduce crime by manipulating perceived sanction risk might want to look far beyond objective risk and consider the factors that actually affect the way in which people estimate the likelihood that they will be apprehended.
Reference: Pogarsky, Greg, Sean Patrick Roche, and Justin Pickett (2017). Heuristics and Biases, Rational Choice, and Sanction Perceptions. Criminology, 55(1) 85-111.
Publication Type
- Regular Issue