View this issue as a PDF: CrimHighlightsV18N1.pdf
This issue of Criminological Highlights addresses the following questions:
- Does the expungement of criminal records make sense?
- What predicts immigrants’ assessments of the police?
- Can policing affect the offending rate of young boys?
- What is the effect of police stops on school performance?
- Did a reduction in active policing cause an increase in homicides in the US?
- How can penitentiaries avoid the use of administrative segregation?
- Does adolescent victimization have similar effects for all types of youths?
- Are sex offender registration and notification laws based on good evidence?
Canadian data (Criminological Highlights 15(2)#6) suggest that the majority of those applying for pardons in Canada do so for reasons related to employment. This study examines the actual effects of the sealing of criminal records in Michigan on the employment of those whose records were sealed.
At least 36 of the 50 US states have provisions that allow records of adult convictions to be sealed, set aside, or expunged. The laws generally give the person the right to answer “No” to questions about criminal records from employers and landlords. The details of these provisions (e.g., in terms of which offences are covered by the legislation) vary widely as do the waiting periods. In some cases, employers or landlords who might know, independently, of a (pardoned) criminal conviction, are prohibited from using that information in their decisions. Perhaps the most interesting statute is Pennsylvania’s 2018 law, which received broad legislative and citizen support. It created an automatic computerized process for sealing certain eligible records after a specified period of time without a subsequent conviction. However, it applies only to minor offences.
Between 1983 and 2011, Michigan’s law allowed applicants to apply to have a single criminal conviction ‘set aside’ after 5 crime-free years. Though the breadth of crimes covered by the law is wide (including most violent felonies), only people with one conviction were allowed to apply. About 5%-6% of eligible people applied for and received pardons within 5 years of eligibility (10 years after conviction) under this law. This is not dramatically different from Canadian experience where it is estimated that around 10% of those with criminal records have received pardons (at any point in their lives). The reasons for the low uptake rate in Michigan appear to relate to six factors: lack of information about the possibility of getting one’s record ‘set aside’, the enormous administrative hassle in gathering the relevant documentation for the application, the $50 fee (Canada’s fee is C$631), distrust of the criminal justice system, lack of access to lawyers to answer questions, and low motivation to remove the criminal record. Uptake of the set-aside provision was highest among women, people convicted when young, those with a non-violent felony, public order, or drug conviction, and those who had not been incarcerated.
By linking criminal record data with employment data, the researchers were able to investigate the relationship between having one’s criminal record ‘set aside’ and the employment and income rates. Those who were employed in the quarter immediately before the application were more likely to apply than those who were not employed. But in addition, those who had suffered at least at 20% wage loss in the previous year were also more likely to apply.
As is the case in Canada, very few people who had their records set aside were subsequently convicted of a new offence (4.2% overall, 1% for a felony) and almost nobody (0.6%) was subsequently convicted of a violent offence. Immediately after having their criminal records set aside, people experienced a jump in employment rates and in wages. It appears that the increase in wages is largely the result of previously minimally employed people working more or getting higher wages.
Conclusion: The findings “strongly support the increased availability of expungement [of criminal records] and particularly efforts to make them automatic, or at least procedurally easy to obtain” (p. 53). Expungement of records has no identified risks to public safety and clearly it has personal and societal benefits in terms of employment. Said differently, removing, as much as possible, the stigma of having a criminal conviction appears to have value both for the person with a record of offending and society more generally.
Reference: Prescott, J.J. and Sonja B. Starr (2019) Expungement of Criminal Convictions: An Empirical Study. Harvard Law Review 133 (in press).
In some countries, such as Canada and the United Kingdom, immigrants hold more favourable views of the police than do native-born residents. Canadian data demonstrate that this is largely due to the views held by recent immigrants from non-democratic countries.
Immigrants to some western countries tend to hold more favourable views of the police than do people who were born in the country. There is some evidence, however, that immigrants’ views are based, in part, on their experiences in their country of origin. Such speculation is consistent with the finding that the difference between immigrants’ views and the views of native born is smaller for those who have lived in the host country for a long time.
In 2016, about 22% of Canadian residents were born outside of Canada. Immigrants, like members of Canada’s “visible minority” communities, do not constitute a homogeneous group. They vary on many dimensions that may be relevant to understanding their views of the police. One important dimension may be their countries of origin. It would be reasonable to expect that immigrants from fully democratic countries would have had very different experiences with the police in their countries of origin than immigrants from countries that do not have stable democratic traditions. Assuming that after arriving in Canada, they view Canadian police from the same perspective as they viewed police in their countries of origin, it can be hypothesized that country of origin should be related to their views of Canadian police.
This paper looks at Canadians’ views of police as a function of their country of origin and the length of time that they had been in Canada. The analysis focused on urban Canadians because almost 96% of immigrants in Canada live in urban areas. Data were obtained from Statistics Canada’s 2014 General Social Survey which contained 6 questions evaluating police performance and a seventh question on confidence the respondent had in the police. The 7 questions, which were highly inter-correlated, were combined into a single scale of favourability toward the police. Immigrants’ country of origin was used to divide them, using the Economic Intelligence Unit’s Democracy Index, into two groups: survey respondents from “full democracies” and respondents from flawed/authoritarian/mixed regimes. In addition, respondents were divided into those who had been in Canada for at least 10 years and those who had immigrated more recently.
The four immigrant groups (divided by type of country of origin – full democracies vs. others – and recent immigrants vs. long term residents of Canada) were compared to native born Canadians. Recent immigrants from full democracies did not differ from native born Canadians in their views of police. The group that was, by far, the most favourable toward the police and significantly different from all other groups was recent immigrants from countries that do not qualify as full democracies. With additional time in Canada, however, immigrants from non-democratic countries became indistinguishable from those who had originally come from democratic countries. The difference between recent immigrants from non-democratic countries and the other 3 groups held even after controlling for age, sex, marital status, education, visible minority status, perception of neighbourhood crime, contact with the police, and whether the person reported having experienced discrimination.
Conclusion: Recent immigrants from non-democratic states may be evaluating the police largely in contrast to what they are used to in their countries of origins. Not surprisingly, then, they evaluate Canadian police more favourably than do recent immigrants from democratic countries. Over time, the difference that country of origin makes diminishes and immigrants’ views begin to look much more like those of respondents who were born in Canada.
Reference: Jung, Maria, Jane B. Sprott and Carolyn Greene (2019). Immigrant Perceptions of the Police: The Role of Country of Origin and Length of Settlement. British Journal of Criminology (in press)
Police actions can increase crime rates: Routine police stops of Black and Latino boys increase the likelihood that these boys will commit offences in the future. These stops also lead to boys reporting more psychological distress.
There is growing evidence that there are negative effects of police stops on young people (e.g., Criminological Highlights 17(4)#6; Understanding the Impact of Police Stops – on the Criminological Highlights “collections” website). This paper looks at the impact of police stops, questioning, and (sometimes) frisks of Black, Latino and other non-white boys – who are comparably more likely (Criminological Highlights, 16(3)#4) to be stopped by the police than are white boys.
This study collected data from boys (23% Black, 58% Latino, 19% other non-white) in Grade 9. They were then interviewed every 6 months for two years. Multiple waves of data collection meant it was possible to examine changes in the behaviour of the boys after they had been stopped (or not) since the previous time period. The boys were recruited from 6 high schools in a large southern US city. They were asked whether, when out walking, they had been stopped by the police (and frisked or not). Youths were also asked about delinquency in the previous 6 months using a 15-item set of questions about specific forms of delinquency. Finally psychological distress was measured with questions about the frequency in the previous 6 months of experiencing negative emotions (e.g., “I feel downhearted and blue”; “I felt I was close to panic”).
There were no important (or consistent) differences in the findings across the three racial groups. About 40% of the boys were stopped (and sometimes frisked) at least once during the 2 year period of the study. Experiencing one or more stops by the police was associated with greater subsequent involvement in delinquent behavior for all three racial groups. The relationship between police stops and delinquent behaviour 6 months later was stronger for younger adolescents than it was for older adolescents. Moreover, delinquency at one stage of the study did not predict subsequent police stops, again supporting the inference that the relationship between police stops and delinquent behaviour is not a result of delinquent youths being more likely to be stopped than non-delinquent youths. Instead, the results support the conclusion that stops cause increased delinquency.
In addition, “adolescents who had experienced more frequent police stops reported greater concurrent distress and greater concurrent delinquency” (p. 8265). Equally important is the finding that “the more police stops a boy experienced, the more distress he reported six months later” (p.8265-6). Once again, distress at one wave did not predict police stops 6 months later, supporting the inference that there is a causal link between police stops and psychological distress. Furthermore, distress – in part the result of police stops – did predict greater delinquency 6 months later.
The conclusion that there is a causal link between police stops and subsequent delinquency is strengthened by the fact “at each wave of the survey, boys who reported little or no involvement in delinquency at the prior wave were just as likely to have been stopped by police 6 months later as boys who had reported higher levels of delinquent behaviour at the previous wave” (p. 8266).
Conclusion: In this study of Latino, Black, and other non-white boys, youths who were stopped by the police showed higher subsequent levels of delinquency than youths who had not been stopped. In addition, prior law-abiding behaviour by a youth did not protect the youth from future police stops. Hence there was apparently little a boy could do to avoid being in jeopardy of being stopped. Stops appeared to cause both increases in psychological distress and increases in delinquency. The “findings suggest that the single most common proactive policing strategy – directing officers to make contact with… boys… in ‘high-crime’ areas – may impose a terrible cost” (p. 8367).
Reference: Del Toro, Juan (and 9 others) (2019). The Criminogenic and Psychological Effects of Police Stops on Adolescent Black and Latino Boys. Proceedings of the National Academy of Sciences 116(17), 8261-8268.
Aggressive policing – saturating high-crime areas with police who are engaged in aggressive order-maintenance policing – interferes with the educational success of African-American boys.
Research has shown that aggressive encounters between the police and Black city residents can have negative impacts on the mental health of Blacks who live nearby (Criminological Highlights 17(5)#2). This paper looks at another important outcome: the school achievement of African-Americans and Hispanics who were exposed to high rates of aggressive policing in New York City.
Between 2004 and 2012, the New York City police ordered, as part of its “Operation Impact,” about 1500 new police recruits to engage in aggressive order-maintenance policing, targeting disorderly behaviour in designated neighbourhoods. The deployment pattern was such that every six months there could be a change in whether a given neighbourhood received this program. On average, high impact zones remained as such for 12.3 months (range 5.3 months to 7.5 years). Rookie officers working in these zones were encouraged to conduct high volumes of investigative stops and to arrest people for low-level offences (e.g., open containers of alcohol).
School records were examined during this period for all public schools in New York for students age 9 to 15. These records contained not only measures of school performance, but also data on race, gender, age, eligibility for free lunch programs (a proxy for the economic status of the child) and whether the student was a new English speaker. The police data used in the study involved records from the “Stop, Question and Frisk” program in the city.
From the police data, it was possible to identify the ‘high impact’ zones. From the education records, it was possible to determine if and when a student lived in one of those high impact zones and, therefore, whether the student had been exposed – directly or vicariously – to high impact aggressive policing. The main dependent variables that were used were state-wide standardized tests of English and Math. These tests were taken in the spring of each year by all students in Grades 3 through 8. The independent variable was the number of days in the previous year that the neighbourhood had been exposed to the high impact policing. The analyses were restricted to African-American and Hispanic students because too few White and Asian students lived in high impact zones. Though the pattern of results was similar for Math, the findings on the impact of policing on the standardized English Language Arts tests were stronger.
For young (age 9) African-American boys, there was no apparent impact of the policing strategy on the English Language Arts scores. However, as this group grew older, high impact policing had the effect of lowering test scores such that by age 12, the effect was modest in size and statistically significant. By age 15, the impact was substantial. In fact, high impact policing lowered test scores of African-American boys by about the same amount as some of the most popular intervention programs would be expected to increase educational achievement. The program also reduced school attendance. These results “suggest that policing programs such as Operation Impact can wipe out the potential benefits of other costly [educational] interventions” (p. 232). From other analyses, it appears that the harmful effects of Operation Impact were evident during the time that it was in place but dissipated somewhat afterwards. Nevertheless, after two years, the effect was still statistically significant. There was some evidence that the program may have reduced violent crime rates, though this impact disappeared as soon as the program ended. There was no measurable impact on African-American girls or on Hispanic children.
Conclusion: The findings demonstrate that aggressive policing strategies can negatively affect the educational prospects of those most exposed to them – in this case by lowering school performance of the group – African-American boys – who were most directly affected by the strategy. Hence “policing strategies and tactics may perpetuate racial inequalities in educational outcomes” (p. 239).
Reference: Legewie, Joscha and Jeffrey Fagan (2019). Aggressive Policing and the Educational Performance of Minority Youth. American Sociological Review, 84(2), 220-247.
Homicide rates in the US increased in 2015 for the first time in many years. There is no reliable evidence that this increase was caused by the so-called “Ferguson effect” – a reduction in proactive police enforcement or a reduction in arrests after a number of highly publicized police killings of unarmed civilians in the US in 2014.
Between 2014 and 2015, the US homicide rate increased by 11.4%, the largest one year increase since 1968. The increase in large cities was even greater (17.2%). The rates remained high through 2017. In 2014 and 2015, there were a number of police killings of unarmed civilians, one of the most famous being the shooting of Michael Brown in Ferguson, Missouri. Some commentators have suggested that changes in police tactics after these events led to an increase in homicides.
In the absence of any evidence, the term “Ferguson effect” was coined, suggesting that the criticism of the police for killing unarmed civilians led police to disengage from proactive enforcement which, in turn, was seen as the cause of the homicide rate increase. Other research (Criminological Highlights 17(3)#4) has suggested that the increases in homicides seen in some cities reflect the kind of variation that one could expect by chance. This paper looks directly at the hypothesized cause of the increase in homicides — the rate of arrests by the police in 53 large US cities — and examines systematically whether changes in arrest rates might explain the increased number of homicides.
Arrest rates, and changes in arrest rates over time, vary across cities. If there really is a ‘Ferguson effect’ that accounts for the increase in homicides, one would expect, controlling for other possible causes of homicides, that decreases in arrest rates would be associated, within cities, with increases in homicides.
This paper looks at the effect of arrest rates for five categories of offence for Blacks and for Whites between 2010 and 2015. In addition, three control variables were incorporated into the model: the percent Black in the city, a measure of economic disadvantage and the rate of Hispanic immigration into the city. The last measure was included because of “assertions of President Trump and others that Hispanic immigration increases the level of violent crime,” an assertion that, not surprisingly, is exactly the opposite of what has frequently been found (Criminological Highlights 17(4)#1, 17(1)#3, 16(1)#2, 14(1)#8, 13(6)#7, 11(1)#4, 10(6)#7, 5(4)#6). Statistical methods were used that attempted to address the fact that it is difficult, in situations such as these, to determine whether something like changes in homicide rates are a cause or an effect of changes in arrest rates.
The major problem with the “arrest rate decline” explanation of the increase in homicide rates is that “arrest rates [for Blacks and Whites] had been dropping for several years before the controversial police killings in Ferguson and elsewhere…. There is also no evidence of a sharper drop in arrest rates in 2015 than in earlier years as would be expected….” (p. 60). Said differently, the drop in arrest rates predated the police killings of civilians in Ferguson. In more sophisticated analyses, the total, Black, and White arrest rates did not have a consistent relationship with homicide rates. In fact, in some instances higher arrest rates were associated with increases in homicide rates.
Conclusion: The investigation, then, looked at regression models of homicide rates, arrest rates and control variables in 53 large cities. The models “allow for the… mutual influence of the arrest and homicide rates in each year after 2010” (p. 67) but provide little support for the notion that ‘de-policing’ was the cause of the increase in homicide rates.
Reference: Rosenfeld, Richard and Joel Wallman (2019). Did De-Policing Cause the Increase in Homicide Rates? Criminology and Public Policy, 18, 51-75.
Canadian penitentiaries are able, if they wish, to identify, at admission, those who are highly likely to end up in administrative segregation. The goal of reducing the use of segregation could be accomplished, in part, by providing appropriate interventions or programming to those who, without intervention, have a very high likelihood of ending up in segregation.
Close to a quarter of Canadian prisoners in federal penitentiaries spend at least six consecutive days in administrative segregation within two years of admission. Though it is broadly agreed that the use of segregation (also referred to as solitary confinement) should be minimized, most of the focus of these efforts is on the decision-making process that leads to a person being put in segregation or being released back to the general population.
This paper presents data related to an alternative approach to reducing the use of segregation: the development of a “Risk of Administrative Segregation Tool” designed to identify, easily and reliably, at the time that a prisoner is first admitted to penitentiary, whether that person, without intervention, is likely to end up for at least 6 days in administrative segregation. The research was first carried out on a sample of 11,110 prisoners and then validated on a different group of 5,591 prisoners. Over 400 potential predictor variables were examined.
Although a large number of factors predict the use of segregation, a scale will be most useful if it can be constructed using variables that are readily available at or shortly after admission to penitentiary. Many of those who are placed in segregation (22%) have their first segregation placement within 60 days of admission (10% within one week).
Various scales, including standard ‘risk’ scales were examined. The variable being predicted was whether the prisoner was placed in segregation (for at least six days) for “the inmate’s own safety” or for “jeopardizing security of the institution” (p. 5) within two years of admission. Data suggest that about 90% of prisoners who are ever placed in segregation are placed there within two years of admission.
In developing the scale to predict segregation, a number of guiding principles were used. Simple options were preferred over complex measures if they didn’t result in meaningful loss of accuracy. Item weights were kept simple. Information had to be readily and ordinarily available at admission. The study explored whether it was preferable to develop four separate scales for Indigenous and Non-Indigenous men and women. Initially both static variables and dynamic variables were examined, but in the end, a single scale composed of static elements was chosen because its accuracy was essentially as good – for all four groups – as that of the more complex scales.
The final scale that was chosen had six items (age, number of prior convictions, previous admissions to administrative segregation, sentence length, number of different types of offences of convictions, and prior convictions for violence). Points were assigned for each item. Overall scale scores ranged from 0 to 12. The highest risk 22% of the prisoners on the scale had a 56% chance of being placed in segregation. If an intervention or treatment program could be developed that would be effective in eliminating the use of segregation for only half of this relatively small group, the use of administrative segregation would, overall, immediately be reduced by about 25%.
Conclusion: This study did not explore what interventions might be carried out early in a prison sentence to reduce the use of segregation of ‘high risk for segregation’ individuals identified at admission. Given that those most at risk of segregation placement can be reasonably accurately identified, it would appear that preventing the need for placement in segregation should be explored as a way of reducing the need for it. At the same time, it is clear that “the scale should not be used to preemptively segregate” or to label prisoners. Instead, this tool could be used to focus resources on reducing the use of segregation, thus freeing up other resources for more positive purposes.
Reference: Helmus, L. Maaike, Sara Johnson, and Andrew J. R. Harris. (2019). Developing and Validating a Tool to Predict Placements in Administrative Segregation: Predictive Accuracy with Inmates, Including Indigenous and Female Inmates. Psychology, Public Policy and Law (online first).
The harmful effects of adolescent violent victimization on adult outcomes are more pronounced among youths whose backgrounds would suggest that they were the least likely to be victimized. For the most disadvantaged youths, the added negative impact of violent victimization did not add to their already relatively high likelihood of problematic life outcomes.
It is well established that experiencing violent victimization as a youth is not evenly distributed across the population. Hence it is not surprising that, in general, people who are likely to have experienced violent victimization as youths are more likely to experience negative life outcomes.
Previous research has demonstrated that youths who have been victimized tend to complete fewer years in school, are at greater risk of depression and anxiety and tend to be more likely to experience violence as adults. What is less clear, however, is whether experiencing violent victimization causes these negative adult outcomes, or if it is only associated.
This study followed youths starting in Grades 7-12 for about 7 years until they were 18-26 years old. In the second year of the study, youths were asked if they had experienced a serious violent victimization in the previous year (i.e., had a knife pulled on them, had been stabbed, shot at, or jumped).
Using data on 50 different variables collected at the first wave of data collection, each youth’s likelihood of being violently victimized was estimated. One year later, 15% of the youths had, in fact, been victimized. Those who were victimized in this first year were matched with youths who had not been victimized but, in fact, had essentially the same likelihood of being victimized (based on the earlier data). Hence, on relevant dimensions, the ‘victimized’ and ‘not victimized’ groups were equivalent. Seven years later, when the participants were age 18-26, measures were obtained on their violent and property offending and violent victimization in the previous year, their hard drug use in the previous month, the experience of depression in the week immediately before being interviewed, and their educational attainment.
The people who had experienced violent victimization as youths were, in comparison to their matched controls, more likely, as young adults, to report violent offending and victimization and lower educational attainment. The sample was then divided into those who had, as youths, a low, medium, or high likelihood of being victims of violence (based on their circumstances when young). The results were quite consistent: “The consequences of adolescent violent victimization on violent offending, [violent] victimization, depressive symptoms, and low educational attainment in early adulthood are strongest among those with the lowest risks of being violently victimized during adolescence” (p. 121-2).
Conclusion: The results show that the consequences of adolescent violent victimization on outcomes in early adulthood are concentrated most strongly among those with the lowest risk of being victimized. It is not, however, that the high-risk youth are more resilient. These youths “were already predisposed to experience these problematic life outcomes…. Low-risk youth, on the other hand, seem to have been diverted onto a negative life pathway that they were unlikely to have experienced prior to being violently victimized” (p. 125).
Reference: Turanovic, Jillian J. (2019). Heterogeneous effects of adolescent violent victimization on problematic outcomes in early adulthood. Criminology, 57, 105-135.
Sex offence registries and public notification of the presence of sex offenders in a community are “based on myths, false and misguided assumptions” about those who have committed sex offences. The presence of these laws “breeds, facilitates, and justifies the emergence of additional misguided criminal justice policies which further expands the reach of the symbolic purpose of the initial misguided policies” (p. 150).
Public beliefs about sex offenders only rarely have a basis in evidence (see Sex Offenders and Society’s Responses to Them under ‘special issues’ on the Criminological Highlights website). In many countries, including Canada, people convicted of certain sex offences must register their whereabouts with local police. In some locations, the information in these registries is available to ordinary citizens.
Existing data suggest that sex offenders, as a group, are quite heterogeneous. Yet sex offence registration and notification (SORN) laws assume that sex offenders are almost invariably high risk to reoffend, have stable offending rates over time (‘once a sex offender, always a sex offender’) and are fundamentally different from those who commit other kinds of offences. Each of these assumptions have been the focus of a substantial amount of research showing them to be untrue. For example, the data suggest that “if offending did persist among sex offenders, it was generally non-sexual in nature” (p. 136). Similarly, the data do not support the notion that adolescent sex offenders are “on a path of life-course persistent sex offending” (p. 136).
SORN laws, then, do not emerge from a careful analysis of evidence. Instead, they appear to emerge from public outcry in response to high-profile sex crimes (often involving the sexual assault and death of a child). The problem is that “There is no empirical evidence showing that registries facilitate police investigation and help solve sex crimes nor that they increase the probability of detention” (p. 138). One obvious problem is that although the presence of these registries may be known, they are often not accessed by members of the public. The evidence on the relationship of the existence of these registries to the rate of sex offending is, at best, mixed. This is not terribly surprising because even if there were a deterrent impact of the presence of a registry, the studies of the quality of the data contained in them suggests that much of the data (e.g., residence locations), at any given time, is not accurate. But in addition, since a substantial number of those convicted of sex offenders are known to their victims (e.g., family members), one would not expect registries to be terribly effective.
Moreover, there are potential harms of SORN laws on those subject to them, largely in terms of their ability to reintegrate peacefully into the community. Restrictions on residence, for example, may interfere with stable reintegration. And again, these restrictions appear to be based on the false assumption that offences take place near an offender’s residence.
Conclusion: “Research showing that the assumptions that sex offence registration and notification laws make about sex offenders are false has not stopped these laws from spreading across all U.S states” (p. 150). These laws have negative impacts on people who, though they may have once committed a relevant offence, are unlikely to reoffend in the future. These negative impacts (e.g., harassment, victimization, eviction) may affect not only the offenders but their family members.
Reference: Lussier, Patrick and Jeffrey Mathesius. (2019). Trojan Horse Policies: Sexual Predators, SORN Laws and the American Experience. Psychology, Crime & Law, 25(2), 133-156.