View this issue as a PDF: CrimHighlightsV19N2.pdf
This issue of Criminological Highlights addresses the following questions:
- What do we know about the relationship between firearms ownership and fatal violence?
- What kinds of jobs can Black, White, and Hispanic prisoners expect to find in their first year after release from prison?
- Does any kind of employment reduce the likelihood of reoffending for those released from prison?
- How does a period of high imprisonment have an impact on imprisonment rates decades later?
- Does existing research support the investment by police services in body-worn cameras?
- What can happen when police ask a question during the interrogation of a suspect that assumes an unproven fact?
- Does the mandatory arrest of suspects in domestic violence incidents reduce their subsequent offending?
- Why is the criminal justice system in some jurisdictions becoming more punitive toward those convicted of sex offences?
Details of the relationship between “gun ownership” and “crime” matter. For example, the total number of guns owned by people in a community may be less important than the number of households in that community that contain one or more firearms.
In the past 50 years, we have learned a lot about firearms and crime (see Criminological Highlights 2(5)#7, 4(4)#7, 6(5)#1, 14(5)#3, 15(5)#1, 17(3)#2, 17(4)#3, 18(2)#2).
One of the more important facts to consider, when thinking about gun control, is that gun attacks are much more likely to be lethal than knife attacks. Robberies with guns have a death rate about 3 times that of robberies without guns. “Robberies – particularly gun robberies – made street crime much more deadly in the US than in other developed nations” (p. 1364).
Between 1986 and 2015, the US population increased by 34% but the number of handguns introduced into the US increased from about 1.66 million in 1986 to 6.91 million in 2015. Over the same period, however, the murder rate dropped. Even though there are more guns in circulation, the proportion of households with firearms declined (from about 50% in 1976 to about 30% in 2014). Said differently, there are more guns, but the prevalence of households with guns is lower than it was. Prevalence is important because “In large part the supply of guns to offenders involves the diversion of guns from legal commerce and ownership, a process that is facilitated in a community where gun ownership is prevalent” (p. 1372).
These simple facts, then, complicate the suggestion that more guns mean more crime. It depends on where those guns are housed. “It is only when incidence and prevalence of gun ownership are distinguished that survey data can provide a coherent explanation of the link between firearms availability and the use of guns in violence” (p. 1368). It would appear that the availability of guns for crime may be more linked to the prevalence of guns in households than to the number of guns (incidence) in the community generally.
Early research “found that victims were 3 times as likely to die if the robbers used a gun rather than a knife or razor” (p. 1375). It was known 40 years ago that the proportion of robberies with a gun increased directly with the prevalence of gun ownership, although the overall rate of robbery was not discernibly affected. That still is true. Gun availability is important: Across US cities, it can be shown that an increase of 1000 gun robberies would lead to 3 or more times as many additional deaths as an increase of 1000 non-gun robberies. (p. 1375). Similarly, robberies with high caliber handguns are similar to robberies with low caliber handguns except on one dimension: robberies with high caliber handguns are more likely to result in the death of the victim. The old slogan that “Guns don’t kill people, people kill people” is profoundly misleading; guns, unlike other weapons, kill quickly, without much effort or skill, and at a distance. It is far easier to kill someone with a gun than, say, a knife. Gun use in violent crime is much more prevalent in America than in other wealthy nations, which explains much of the observed gap in homicide rates. Even in the US (with an estimated 300 million guns in private hands), the availability of guns to offenders “differs widely across jurisdictions and among differently situated offenders within the same jurisdiction, and that variation is reflected in weapon choice in violent crime” (1383). This suggests that controls are having an effect, even though they are not perfect.
Conclusion: It is unlikely that countries with many guns in circulation will ever be able to eliminate the use of firearms in crime. However, from what is known about firearms (e.g., death rates related to the prevalence of firearms and the caliber of firearms being used in crime), it would be possible to craft policies that would be helpful in reducing the lethal use of firearms. The challenge therefore is more political than it is criminological and is in developing policy that is consistent with what we now know.
Reference: Zimring, Franklin E. (2020) Firearms and Violence in American Life: 50 Years Later. Cook, Philip J. Thinking About Gun Violence. Criminology & Public Policy 19, 1359-1369;1371-1393.
People released from prison show high rates of unemployment. When they do get work, they typically work for poverty-level wages. This is more the case for Blacks and Hispanics than for Whites in this study even after controlling for various background factors related to employability.
It is well established that involvement in the criminal justice system makes it difficult for people to get a job (Criminological Highlights 6(3)#2, 15(1)#7). This paper examines the casual, temporary, “informal” and formal employment of men and women during their first year after being released from prison in the Boston, Massachusetts, area.
Employment rates for those released from prison are typically very low. In this study, (total) employment rates for Blacks never exceeded 50%. The rate for Hispanics was slightly higher. The rate for Whites was higher, but only reached 70% in one month during the 12 months after release from prison. For those who were working, Blacks averaged about $1300/month -- about half of the income for Black workers in the US labour market as a whole. Hispanics did a bit better (about $1500/month), whereas Whites who were working were making about $2500 a month by the end of the year.
In this study, 122 people were interviewed 5 times during the year after their release in order to get a detailed picture of employment and earnings. Part of the purpose in using interview data on earnings (rather than official records) was that it was a way of including temporary and informal (often cash) work.
As already mentioned, Blacks and Hispanics were less likely to be employed and, if employed, made less money than Whites. Whites, in this sample, had more health deficits (chronic pain, drug addiction, mental illness) than Blacks and Hispanics. When controlling for these factors, then, the size of the employment deficit for Blacks and Hispanics increased dramatically. It was estimated that Blacks earned considerably less than observably identical Whites. This is attributable to two factors: lower rates of employment and lower rates of pay if employed.
One possible reason for the race/ethnicity differences may be that Whites were considerably more likely than Blacks and Hispanics to find work through recommendations and referrals from their social networks. The result of this difference was that Whites were more likely than Blacks and Hispanics to find jobs in high paying industries (e.g., construction). Blacks and Hispanics got jobs through family and friends, but these were typically low-paying jobs. Blacks and Hispanics were also more likely than Whites to suggest that it was their criminal record that stood in the way of getting a job.
Perhaps the most disturbing findings relate to race: Whites were more likely to be disadvantaged because of physical disabilities and drug addiction; hence they were less ready for employment and were more socially isolated than Blacks and Hispanics. “Despite these disadvantages, their employment rates were higher and their earnings were nearly double those of formerly incarcerated Blacks and Hispanics…. Higher levels of employment and earnings among Whites are associated with social network connections to relatively well-paying jobs” (p. 1537). But in addition, because Blacks are more likely to have contact with potential employers as a result of a formal job search, the stigma of a criminal record appears to affect them more than it affects whites.
Conclusion: Labour market outcomes for those released from prison are worse among Blacks and Hispanics than among Whites, notwithstanding various health disadvantages experienced by Whites. “The results suggest that high rates of incarceration among Blacks and Hispanics combine with the social conditions of poverty – characterized by social detachment from skilled employment and the stigma of criminality – to produce high rates of joblessness and low-wage employment” (p. 1538).
Reference: Western, Bruce and Catherine Sirois (2019). Racialized Re-entry: Labour Market Inequality After Incarceration. Social Forces 97(4), 1517-1542.
Not any job will reduce the likelihood of reoffending for young people who were before the courts for serious crimes. For some young people, getting a low-quality job may be criminogenic. For certain groups (e.g., Black youths in some labour markets) transitioning to a job that was rated as higher quality may lead to lower involvement in crime.
It is often assumed that getting a job will reduce the likelihood that people at risk of offending will actually offend. The evidence for this is, however, inconsistent. This paper explores the possibility that the effect, if there is one, may depend on the quality of the job as well as the circumstances of the person who gets a job.
Previous research has suggested that providing former prisoners with jobs may reduce reoffending for those over age 27, but may have little impact on younger people (Criminological Highlights 4(3)#6). This paper explores the possibility that the effects of being employed depend, in part, on the quality of the job that a person gets.
This study followed 1,195 adults (age 18-26) for 4 years who had been to court for serious crimes as adolescents (age 14-19). They were interviewed at least annually, and self-reports of crime were collected at each interview. Participants were asked whether they were working, and, if they had a job, they were asked 10 questions about the quality of the job. These included such issues as how satisfied they were with the salary, supervision, advancement, status, colleagues, and workload. These were combined into a single “job satisfaction” measure for each job they had. Other characteristics of the job (income, benefits, whether it was full time, etc.) were also measured.
In general, it would appear, in simple analyses, that those who were working were less likely to be involved in crime. However, when job satisfaction is taken into account, the “work” effect was accounted for completely by the fact that those with high job satisfaction were less involved in crime. Equally, or perhaps, more important, however, is the finding that there was no reduction in crime associated with low job-satisfaction employment. Indeed, compared to not working at all, working in the lowest quality jobs was associated with higher rates of offending. More specifically “transitioning from not working to working in the lowest job satisfaction category is associated in a 30% increased rate of offending.” (p. 716).
Overall, “an improvement in job quality [had] no significant association with offending” (p. 721) – a finding consistent across various sample subgroups such as different levels of criminal propensity, ages (age 18-21 vs 22-26), sexes, or economic states (prior to or during the 2008 recession). For race, however, the effects were different: “An increase in job satisfaction [for those working was] associated with a 21% reduced rate of offending among Blacks” (p. 721), but not for those who were not Black.
Conclusion: It is possible that one of the reasons that often “work doesn’t work” (p. 722) in reducing offending is that “high-risk individuals with a history of criminal involvement often do not obtain the types of jobs required to improve crime related outcomes” (p. 722). The fact that transitioning from no work to work in an unsatisfying setting was associated, if anything, with an increase in offending suggests that those concerned with helping those who have offended find jobs should be cautious about assuming that any job will be better than no job.
Reference: Jaynes, Chae M. (2020) The Relationship between Job Quality and Crime: Examining Heterogeneity in Treatment and Treatment Effect. Journal of Research in Crime and Delinquency 57(6), 693-740.
People who reach young adulthood during a period of high crime and high incarceration rates are likely to contribute significantly to high incarceration rates throughout their lives. The reason is simple: during high crime & punishment periods they acquired serious criminal histories that result in increased punishment throughout their lives.
“National crime rates [in the US] have dropped steadily to levels last seen in the 1960s…, but prison incarceration rates have remained approximately four times the level of the 1960s” (p. 646). This paper suggests that a large part of the continued high rate of US imprisonment is that those who came of age in the high crime & punishment periods in the 1980s and 1990s continued to have a high likelihood of imprisonment throughout their lives.
This study was carried out in the state of North Carolina (NC). NC’s crime rate, which had risen in the 1980s, had, by 2016, returned to what it was in the early 1970s. But incarceration rates were still high. The 1980s, in NC, featured various tough-on-crime policies: “increased law enforcement activities on the streets, especially for low-level drug offences, and the enactment of mandatory sentencing laws for drug possession, trafficking and impaired driving” (p. 650). Though there was some “early” release (to avoid over-crowding), NC judges responded to the prospect of “early release” by giving longer prison sentences. In 1994, legislation came into effect mandating that the offender’s criminal history and the offence were to be the only determinants of the sentence.
Comparing groups of people who turned 24 during different periods allows one to see that those who “came of age during the crime-punishment wave have elevated incarceration rates throughout their lives” (p. 659). The data also show that people who turned 24 years old (an estimate of the age of highest criminal activity) between 1985 and 2005 had historically high likelihoods of incarceration throughout their observable lifetimes. Said differently, coming of (high-crime-involvement) age during an especially punitive period increases the likelihood that a person will be incarcerated throughout their lifetime.
The data that most strongly suggest that it is the accumulation of a criminal record that has led to continued high rates of imprisonment in NC is the finding that, controlling for offence severity, those who came of age during the high-crime-punishment period have a significantly higher likelihood of receiving a prison sentence if convicted when they are older compared to those who came of age before the high-crime-punishment era. Most importantly, when criminal history is controlled, those who turned 24 during the high crime & imprisonment period were no more likely to receive a sentence of incarceration than those who came of age outside of this period. Criminal history, then “is the primary reason behind the elevated chance of a prison sentence given a felony conviction [for] those who turned 24 during the high-crime-and-punishment period” (p. 662).
Conclusion: The crime-punishment wave during the 1980s and 1990s did, of course, generate increased incarceration rates during that period. But its largest and most long-lasting impact appears to be that the impact of acquiring a criminal record at that time persists throughout the person’s lifetime. Hence the impact of a short-term period of increased crime and/or punishment can last for several decades because of its lasting impact on a person’s criminal record. Looking at the other side of the coin, however, the findings in this paper remind us that policies designed to reduce court involvement and the likelihood of acquiring a criminal record can have lasting effects in reducing imprisonment rates (see Criminological Highlights 18(3)#2).
Reference: Shen, Yinzhi, Shawn D. Bushway, Lucy C. Sorensen & Herbert L. Smith (2020). Locking Up My Generation: Cohort Differences in Prison Spells Over the Life Course. Criminology 58, 645-677.
The effects of requiring the police to use body-worn cameras on both police and citizen behaviour vary enough to make it impossible for any police service to conclude, without its own careful research, that the investment in this technology will have any beneficial effects.
Requiring police to wear and use body-worn cameras (BWCs) is often seen as a method of improving the quality of police-citizen interactions and police accountability. This paper reviews the research on the impact of this new technology (see also Criminological Highlights 15(6)#8, 16(2)#3, 16(6)#2). .
“BWCs are likely the most rapidly diffusing technologies in modern police history” (p. 3). The technology was designed to document interactions between police and citizens so that ‘objective’ records of problematic interactions could be examined. The idea also is that the recording of interactions will deter wrongdoing or undesirable behaviour because a filmed record will remain.
This review first examined 549 documents related to the use of BWCs. Within these 549 reports, only 30 adequately executed studies were found, 10 of which were from one research team. Only studies using experimental (20 studies) or quasi-experimental designs (10 studies) were included. Two general issues were examined: (1) “The impacts of BWCs on officer behaviour, as measured by complaints against officers, officer use of force, arrest and citation behaviour” etc., and (2) “The impacts of BWCs on civilian behaviour as measured by community members’ compliance with police commands” (p. 7). One problem with many of the studies is ‘contamination’ between treatment (BWC) and control (no-BWC). This can happen, for example, when officers are randomly (or otherwise) assigned to wear a camera on some shifts and not others.
Across the 30 studies, 12 outcome constructs were identified (e.g., use of force, complaints against police, field interviews/stop-and-frisk, traffic stops, non-traffic citations). For 7 of these 12 outcomes there were a reasonable number of studies -- 5 or more studies with results on the outcome. For only 1 of these 7 outcomes was there an overall effect: there was a 17% reduction in “complaints against officers” when BWCs were being used. Though the effect was significant overall, it is important to note that in 8 of the 22 studies that examined this construct there was either no reduction or an apparent increase in complaints associated with the use of BWCs. The only other significant overall finding – an overall increase in non-traffic citations – was based on only two studies and hence it is risky to draw firm conclusions about it.
“Use of force” was the most frequently examined construct (26 of the 30 studies). There was no overall effect of BWCs on use of force. Indeed, the most important finding was heterogeneity across studies, with BWCs associated with decreased use of force in some studies and increased use of force in others.
This “meta-analysis of 30 studies and 116 effects of police use of BWCs finds that this technology produces few clear or consistent impacts on police or citizen behaviours. Across a variety of outcome measures – including police use of force, complaints against officers, arrests, proactive police activities, assaults or resistance against officers, citizen calls for police service, and others, individual studies have produced a mix of positive, negative, and null findings” (p. 31). Though there does appear to be an overall reduction in complaints against the police, “It is unclear… to what extent this represents improvements in the behaviours of officers and citizens toward one another… or a decline in the willingness of citizens to file complaints against officers” (p. 31).
Conclusion: Given the variability across studies, and the fact that most of the studies used ‘true experiments’ providing rather reliable evidence, “if law enforcement agencies continue to acquire BWCs, they and their communities should temper their expectations about them. This systematic review does not provide strong support that BWCs have the impacts on certain outcomes that were initially expected” (p. 34).
Reference: Lum, Cynthia & 7 others (2020). Body-worn cameras’ effects on police officers and citizen behaviour: A systematic review. Campbell Systematic Reviews 16(3), September 2020.
When the police at a formal interrogation ask accused people questions about incriminating evidence that doesn’t exist, those listening to the interrogation are likely to believe that the accused is guilty.
One of the most influential textbooks on police interrogations encourages police to use “a ‘bait question’ [which] involves asking about hypothetical evidence that may not exist” (p. 902). Though little is known about whether this approach is useful as a lie detection tool, this paper adds to our knowledge of an important related matter: Does the use of bait questions “cause memory distortion for what evidence really does and does not exist”?(p. 902).
Previous research has suggested that “bait questions supply misleading post-event information that [those hearing about it] then incorporate into their memory for the case report” (p. 904). This paper takes it one step further: Could this memory distortion affect the outcome of a criminal trial by increasing the perception of guilt by decision makers?
The paper presents three experiments that attempt not only to see whether those hearing ‘bait questions’ being asked are likely to see the accused as guilty, but also to try to understand why this happens. The structure of the experiments was quite similar. A hypothetical crime (a murder or a robbery) was described. The manner in which the police identified the prime suspect was explained and key pieces of evidence (six in the first experiment) were presented. Participants read a (hypothetical) transcript of the police interrogation of the prime suspect in which, for some study participants, a ‘bait question’ was included: “Is there any reason we would find a witness who saw you leaving the docks that night?” (p. 908). The bait question implied a fact – that the accused was leaving the docks that night – for which there was no evidence. The suspect in the case rejected the evidence as impossible.
The first experiment was designed to see whether the effect of the ‘bait question’ varied across types of cases. The cases varied on whether the misleading evidence was strong or only circumstantial evidence. And there were either two or four misleading (bait) questions. The result suggested that these differences had no effect on the verdicts that study participants offered. However, in all cases, where the police asked at least one ‘bait’ question (implying that the accused was guilty), the study participants were more likely to view the accused as being guilty. Without any bait questions 29% of the respondents thought that the accused should be found guilty. With any bait questions (implying strong or circumstantial evidence), the accused was more likely to be seen as guilty (38% to 47% saw him as guilty, across the 4 conditions). The second and third experiments tried, unsuccessfully, to determine how misleading hypothetical evidence affects case outcome.
Taken as a whole, it is suggested that “ambiguous, nonprobative evidence can inflate guilt perception” (p. 920). About a third of the subjects in the second and third experiments who were exposed to ‘bait questions’ incorrectly remembered the evidence as having been collected by the police. These respondents, not surprisingly, were more likely to view the defendant as being guilty.
Conclusion: Hearing that the police, when interrogating a suspect, asked a question implying the existence of incriminating evidence that they did not have, appears to increase the likelihood that an accused person would be seen as guilty even when the evidence was never actually introduced. The findings from this study “strongly suggest that jurors may not differentiate [between] the real evidence heard in trial and the non-existent bait question evidence in memory” (p. 921). “Bait questions, then, distort memory for what evidence actually exists” (p. 922).
Reference: Crozier, William E., Timothy J. Luke, and Deryn Strange (2020). Taking the Bait: Interrogation Questions about Hypothetical Evidence May Inflate Perceptions of Guilt. Psychology, Crime & Law 26(9), 902-925.
A systematic review of the practice of mandatory arrest for domestic violence concludes that arrest does not reduce the likelihood that another offence will occur.
Prior to the 1980s, domestic violence was often seen, in many places, as being a private matter that did not necessarily justify police involvement. In 1984, a highly publicized, but somewhat flawed, experiment was published that suggested that ‘mandatory arrest’ by the police of men who assaulted their spouses would reduce similar re-offending. This paper examines the research related to this suggestion.
Given the range of circumstances of those involved in crime, the range of severity of the assaults, and the range of alternatives approaches that might be available, mandatory responses to crime can have serious unanticipated consequences (see Criminological Highlights 19(1)#1). Nevertheless, mandatory arrest policies are common (28 US states have mandatory/preferred arrest statutes for domestic violence).
The first experiment on mandatory arrest was published in 1984 and between then and 2015, there were 10 additional attempts to examine the impact of mandatory arrest. In all cases, immediate arrest (or the issuing of an arrest warrant) was compared to some other treatment (no arrest, mediation, separation, a protective order, counselling, etc.).
In the analyses carried out for this paper, if a study contrasted ‘arrest’ with two or more alternative treatments (e.g., mediation or a protection order), these multiple comparison groups were combined into a single “no arrest” comparison group. In other words, this paper contrasts “arrest” with “something else” (which varied across studies). Repeat offending was operationalized in a variety of different ways. In some studies, any new offence was counted as a repeat offence. In others, calls for service by or on behalf of the original victim for a new domestic violence incident was used as the outcome variable.
The fact that this meta-analysis relies on official records (rather than reports of victims) is obviously a limitation on the findings since it is possible that certain treatments (arrest, mediation, counselling, for example) may affect the likelihood that a domestic violence victim will report a subsequent offence by the same offender to the police. However, in cases in which both official records and victimization reports were included in the original paper, the results tended to be the same.
Looking at the 11 studies as a group, there was no overall impact of mandatory arrest on subsequent domestic violence or offending more generally. There was, however, some variability across studies suggesting that there may be specific circumstances/ locations/ procedures that are used by the police that are effective even though looking across studies, the “mandatory arrest” of domestic violence suspects as a broad construct was ineffective in reducing subsequent offending.
Conclusion: “The results of the meta-analysis (of the 11 studies identified in this paper) suggest that the mandatory arrest policies adopted by police departments in the 1980s and 1990s did not produce the desired deterrent effect sought by law enforcement agencies and advocates alike. That is, arrest did not have a significant effect in reducing the likelihood of repeat offending among individuals arrested for domestic violence” (p. 7). It would seem that the burden of proof to demonstrate ‘effectiveness” should be on those advocating or implementing ‘mandatory arrest’ policies.
Reference: Hoppe, Susan J., Yan Zhang, Brittany E. Hayes, and Matthew A. Bills (2020). Mandatory Arrest for Domestic Violence and Repeat Offending: A Meta-analysis. Aggression and Violent Behavior 30, 1-9.
Increasing levels of public concern about sex offenders – fed largely in the US by national media – can have the effect of increasing the punishments handed down by the courts even though there is no evidence of increasing rates of sex crimes or increased seriousness of sex offense cases that come to the attention of police.
There are many myths about sex offenders, most notably that the likelihood of the repetition of a sex offence by a person once convicted of a sex offence is high (“once a sex offender, always a sex offender”). This and various other myths have been the subject of a substantial amount of research (see the Criminological Highlights ‘special issue’ on our website).
“Moral panics” about certain crimes are not unusual, though the focus of them does vary from place to place and time to time. [Recent Canadian governments, for example, have made four separate changes in the “bestiality” sections of the Criminal Code and have increased the penalties for harm to certain animals, notwithstanding the fact that there was no evidence of increased problems in this area.]
This paper first looks at a simple question: Were those convicted of sex offences in recent years the target of special increases in punishment? If convicted sex offenders were targeted for increased punishment, what seemed to be driving this change?
In the 1980s and 1990s, public concern in the US about sex offenders seemed to increase. From a policy perspective, the preferred solution appeared to be incapacitation. Support for tougher sanctions appeared higher among those who believed that the incidence of these crimes was increasing. This study uses Florida state court data to illustrate how sentencing for sex offenders has changed. Even though Florida sentencing guidelines did not change during the period examined (1995-2011) and more importantly, the “offence severity” of these offences are fixed in the statute, judges have the discretion to change the severity of sentences by departing from guidelines. In this study, changes in the severity of the courts’ responses to sex offences were compared to changes (if any) in courts’ responses to other offences.
Between 1995 and 2011, the likelihood that a sex offender would be imprisoned increased substantially more than the increase in the likelihood of incarceration for an offender convicted of other violent, drug, property or other offences. Similarly, sentence length increased more dramatically for sex offenders than for any of these other groups of offenders. “Changes in sanctioning practices during the study period cannot be attributed to changes in the average severity of sex offences or the composition of sex offences….” (p. 105). In other words, the changes in sentencing did not reflect changes in the offences: the average severity of sex offences was stable. It was only punishment for them that increased.
During the period that was studied, rates of sexual violence in Florida and in the US (overall) declined. At the same time, the number of headlines in national newspapers concerning sex offenders increased substantially, though there was a slight decline, particularly between 1995 and 1998, in headlines involving sex offences in the Florida newspapers.
Conclusion: The increase in punitiveness (the likelihood being imprisoned and length of the prison sentence if a person was incarcerated) that occurred in Florida in the case of sex offences appeared to the result of increased social concerns about sex crimes, not the severity or prevalence of these offences. “Under the assumption that [the increase in media attention to these crimes] operates as a proxy measure of social concern about sex crimes, [the study] found evidence that increasing social concern about sexual violence at the national level is positively associated with, and explains some of the punitiveness toward, sex offenders in Florida state courts” (p. 105-6). More generally, the findings suggest that those responsible for punitive responses to a particular type of offence should consider carefully what the reasons might be for changes – and in particular increases – in punitive responses.
Reference: Cochran, Joshua C., Elisa L. Toman, Ryan T. Shield, and Daniel P. Mears (2021). A Uniquely Punitive Turn? Sex Offenders and the Persistence of Punitive Sanctioning. Journal of Research in Crime & Delinquency 58(1), 74-118.