Criminological Highlights Vol. 22, No. 3 - June 2025

Item 1

Carefully executed experiments with representative samples of American citizens suggest that there is substantial support for systems of “second look sentencing” where people who have been incarcerated for a very long time can apply to a court to be “resentenced” by a judge.

Imprisonment is seen as an easy solution to crime in large part because many ordinary citizens are not aware of the failure of harsh sentences to reduce crime (see CrimHighlights.ca “special issues”). This paper demonstrates that once someone has been imprisoned for a long time, there is substantial public support for policies that allow for the “resentencing” of such offenders and their return to the community.

“Second look sentencing” (p. 281) is becoming increasingly popular in many US states. In 2017 in the District of Columbia, youths and young adult offenders who had been imprisoned for 15 years or more were made eligible to petition for release. This paper examines public support for extending second look sentencing to all offenders who have served long terms of imprisonment.

In 2023, representative samples of US adults participated in two experiments in which they were asked whether they would support the release of offenders after they had been imprisoned for some time. In the first experiment, the time that they had been imprisoned was varied (randomly) in the questionnaire as was the description of the age when they committed the offence. In the second experiment other information was systematically varied: the programs the offender had participated in, the view of the prison warden, and whether a victim (or victim’s family) supported resentencing and possible early release.

Characteristics of the case that were varied made no difference. In the first study, 56.2% supported second look sentencing, and 19.1% opposed it, with 24.6% having no opinion. In the second study, there was a very small effect of age: support for re-sentencing was greatest for those who offended when they were youths. In addition, support was highest if the offender had completed rehabilitation programs and had support of the warden and/or victim.

The broad level of support from these surveys is similar to some actual “resentencing” experiences in Canada. Since capital punishment was abolished in Canada in 1976, offenders receive life sentences for murder and are ineligible to apply for parole for 10-25 years. However, until 2011, prisoners who had received sentences including parole ineligibility for 15 years or more could go before a jury to have their case reviewed. The jury could reduce this parole ineligibility period, which then allowed the prisoner to go before the parole board. Until the mid-1990s, 8 of the 12 jurors had to agree to reduce the ineligibility period; in 80% of the hearings, juries did so.   However, in 1997, the law was changed such that the “resentencing” was not available to those convicted of more than one murder. Furthermore, a judge had to determine that there was a “substantial likelihood that the application would succeed” (Criminal Code S.745.61 (1)). The most dramatic change was that the 12-person jury had to be unanimous for the ineligibility period to be reduced. Nevertheless, Canadian citizens, acting as jurors in these cases, actually did reduce the parole ineligibility periods for 73% of the cases they heard. In 2011, a Conservative federal government eliminated this form of “resentencing” for anyone convicted of a murder that took place after 2011.

Conclusion:  It would appear that the public is not opposed to some form of “second look” at very long sentences initially given to people who have committed serious offences. “A majority of Americans are supportive of universal second look sentencing and thus are open to releasing some people currently serving long sentences for serious violent crimes” (p 290). More specifically, survey data from the US, and actual decisions by ordinary citizens in Canada, suggest that after 10-15 years of imprisonment, members of the public are willing to allow a second look at the imprisonment decision and to allow many prisoners to be released.

Reference: Smith, Paula plus 6 others (2025). Public support for Universal Second Look Sentencing, A Research Note. Criminology, 63, 289-293. Public Safety Canada: Corrections and Conditional Release Act Statistical Overview, 1998 and 2021.

Item 2

Risk assessment tools are developed and validated on specific populations at particular points in time. Just as it cannot be assumed that they are equally valid for different groups of people, this paper presents data demonstrating that a measure developed at one point in time is dramatically less accurate for youths who were born, on average, 12 years later.

Predictions are often made about the likelihood that a person caught up in the criminal justice system will commit another offence.  Typically, various characteristics about the person (e.g., characteristics of the person’s history, their family and neighbourhood, and their criminal justice background) are used as predictors. But what if the factors that predict involvement in crime change over time?

This paper examines a very simple hypothesis: the accuracy of the prediction from a risk assessment instrument will deteriorate when it is applied to people born a few years later than the original validation sample. The study uses data from 1,057 youths. A “classic” set of variables (e.g., characteristics of the youth, family, poverty, community environment) and a second “full set” of predictors (with 35 measures overall) were used to create prediction measures. As it turns out, the pattern of results for the full model is similar to that of the classic model and the added value of the “full” model was negligible. The outcome measure of interest was whether youths were arrested between the ages of 17 and 24. 

For the older cohort the prediction of arrest was as good or better than the predictions made by risk assessment instruments used daily in criminal justice settings. For the older cohort – the youths on whom the scale was developed – the scale was well calibrated: “On average, the predicted probability of arrest from the model is the same as the proportion of participants who actually are arrested at that probability level.” For example, for scores where the model predicted that 75% of the youths would be arrested, it was found that about 75% actually were actually arrested. For the younger cohort, however, the measure substantially overpredicted offending: Youths who were predicted to have a 75% likelihood of being arrested in fact only had about a 40% likelihood of being arrested. 

When the data were divided into three groups by race (White and other, Black, Latino), the results were very much the same: There was an over-prediction of arrest for the younger cohort but no evidence of over-prediction of arrest for the older cohort (the group that was used for validation purposes) for any of the race groups. When “high risk” youths were examined separately, the same finding appeared: there was over-prediction of arrest for those who were born later than the sample used to develop the measures.

“Societies, just like the individuals who compose them, change over time” (p 6). In this case, the cohort differences that are reported could well reflect changes in the community in which the measure was being used – in this case, an overall decline in arrest in Chicago. The problem, of course, is that in order to determine why the quality of the prediction of future offending decreased, one would have to re-validate the scale.

Conclusion: The results of this study should not be used to question the value of risk assessment instruments generally: Subjective predictions by humans may well suffer from this same problem. “The most straightforward approach to mitigating cohort bias is to ensure that risk assessment instruments are updated frequently” (p. 8). The performance of prediction models is unlikely to be stable over time. The predictions will fail “if the dynamics of social change are ignored.” But in addition, “In criminal justice applications, the costs of such overpredictions of risk level are high” (p. 3) both to the youth and to society more generally.

Reference: Montana, Erika, Daniel S. Nagin, Roland Neil and Robert J. Sampson (2023). Cohort Bias in Predictive Risk Assessment of Future Criminal Justice Involvement. Proceedings of the National Academy of Sciences, 120(23) 1-9.

Item 3

Neighbourhoods with comparatively high rates of violent crime can reduce local violent crime rates by making relatively inexpensive improvements in their physical characteristics.

There is a growing amount of research (e.g., Criminological Highlights 21(6)#6) suggesting that improvements in the physical condition of urban neighbourhoods can reduce local crime. Part of the reason for this may be that depopulated areas of cities “are particularly vulnerable to violent crime because illegal activity can occur in depopulated areas with limited surveillance….” (p. 2).

This study took place in Flint, Michigan, a city that, in the latter part of the 20th century, experienced depopulation and urban decay. In 2011, the city declared a financial state of emergency. In a particularly hard-hit area of the city, federal money was made available to local communities to carry out improvements in the environmental design of their immediate neighbourhoods. These involved improvements in such areas as landscaping, signage, decorative fencing, maintenance of the public areas, removing of graffiti, controls on access to some properties, and the encouragement of positive uses of land (e.g., through the development or improvement of playgrounds). What may have also been important in the City’s plan, however, was that much of these improvements were carried out and/or organized by community residents.

Between 2015 and 2018, improvement activities were carried out in many – but not all – locations in this part of the city. These were typically initiated and carried out by neighbourhood residents. The amount of improvement to the neighbourhood (physical such as maintenance or repairs as well as social such as community events), however, varied dramatically across neighbourhoods. In order to assess the changes that had occurred and the impact of these changes on crime, each “street segment” (both sides of a street between two intersections or the end of the street) in the area was rated such that the “intensity” of the improvements, if any, (between 2015 and 2018) could be scored on 10 different dimensions. In addition, violent crime and violent firearm incidents were obtained from the police for each street segment.

Important controls were included in the analyses: (1) Violence density prior to the start of the improvement activities, (2) Neighbourhood disadvantage (percent of households in poverty, on public assistance, renter occupied, vacant), and (3) Researchers’ ratings of property maintenance (buildings and lawns) at the beginning of the program. 

Looking both at the rates of violent crime in the street segment and violent firearm crime, it was found that “higher levels of community engaged [improvement] activities were associated with steeper declines in violent… and violent firearm density over time” (p. 9).

Conclusion: The concept of “crime prevention through environmental design” is not new. This paper suggests, however, that “active resident participation in improving vacant and deteriorating environments can promote safer communities by reducing opportunities for crime and asserting positive community ownership and oversight” (p. 11). “Community engaged Crime Prevention Through Environmental Design may be particularly beneficial [in reducing violence] because it takes a nonpunitive, population-level approach that avoids some of the costs and harms associated with justice-system interventions” (p. 13). At a minimum, it should be remembered that resources (money and time) spent in this project improved the physical and social quality of the neighbourhoods. Crime reduction effects can be seen as additional benefits to the community from the improvement of the quality of the neighbourhoods.

Reference: Rupp, Laney A. and 9 others (2025). Community-engaged Crime Prevention Through Environmental Design and Reductions in Violent and Firearm Crime. American Journal of Community Psychology, 1-16.

Item 4

Exposure to family violence when a child is young is most likely to be associated with later violent and non-violent crime when the child also experiences residential instability and the incarceration of a parent.

The relationship between childhood family adversity and delinquency has been demonstrated repeatedly in the history of criminology. In this context, childhood exposure to family violence has also frequently been identified as an important risk factor for delinquency.

This study examines whether the impact of exposure to family violence is especially harmful to a child if it is combined with either or both of two other family disadvantages that are known to contribute to an increased likelihood of delinquent acts by a child: residential instability and parental incarceration.  Each of these factors has, on its own, been shown to be associated with delinquency.  It could be argued that these factors – which tend to be correlated with one another – are each an indicator of a difficult family situation for a child. If so, it is possible that experiencing more than one of the three is, in fact, no worse for the child since each may simply be an indicator of a less-than-ideal family. Alternatively, family violence may have an especially large impact when it is experienced by youths who are already vulnerable due to residential instability and the incarceration of a parent.

This paper examines data from 3,191 youths born in 20 large US cities between 1998 and 2000. Mothers – especially unmarried mothers – were recruited for the study in hospitals at the time of the birth of their child. They were interviewed 4 times in the next nine years. When the child was 15, both the mother and the child were interviewed. 

About 14% of the youths experienced exposure to family violence at age 9 or earlier. This was assessed from answers to a question posed to mothers on whether they had ever been “seriously hurt in a fight with their romantic partner” (p. 398) and whether this had occurred in front of the child or when the child was in the home. Residential instability for the youth was defined as having moved 3 or more times when the child was 3-9 years old. The youth was described as experiencing parental incarceration if one or both parents had been incarcerated between the youth’s 3^rd and 9^th birthdays.  33% of the youths experienced residential instability and 32% experienced the incarceration of a parent.  About 14% of the youths experienced exposure to family violence. Various controls – including the delinquency of peers, childhood emotional or physical abuse, as well as maternal education, poverty, violence of the neighbourhood -- were statistically controlled. The main dependent variable was the self-report of the 15-year-old youth of their involvement in 13 separate forms of violent and non-violent offending in the previous year. 

Taken by themselves, “childhood exposure to family violence, residential instability, and parental incarceration are not particularly consistent nor robust predictors of adolescent delinquency” (p. 407) when they are examined as independent predictors. However, when examined in combination with one another, it is clear – especially when looking at any delinquency (violent or non-violent) – that youths who were exposed to all three forms of family adversity (family violence as well as residential instability and parental incarceration) were substantially more likely to be involved in crime.

Conclusion: It appears that the impact of experiencing family violence is intimately linked to other household adversities (residential instability and parental incarceration). Said differently, it is the overall family environment that needs to be understood when trying to understand why a particular youth is responding to household adversities by offending.

Reference: Talaugon, Alyssa R. and Jillian J. Turanovic (2025). Childhood Exposure to Family Violence, Residential Instability, and Parental Incarceration. Journal of Research in Crime and Delinquency, 62, 387-428.

Item 5

English judges sometimes use their judgements of an offender’s remorse, good character, and the ability to rehabilitate in their determination of the sentence that they impose. An examination of the use of these judgements in the sentences imposed on those found guilty of assault or drug offences suggests that the use of these subjective judgments helps Whites more than Blacks.

When imposing a sentence, judges decide whether certain factors are relevant and should affect the sentence. For some factors (e.g., whether an offence was committed when the accused was on bail) the assessment is objective. For others (e.g., whether the defendant shows remorse) the judgement is subjective. This paper examines the possibility that factors where there is high judicial discretion and which relate directly to the offender (rather than the nature of the offence) “have the greatest risk of being ‘racially determined’ and, therefore, operating as a source of ethnic disparity in sentencing outcomes” (p. 242).

This study looks at two high volume offence groups: assault (and public order offences) and drug offences. Using the England & Wales sentencing guidelines, 5 sentencing factors were identified that involve both a high degree of judicial discretion and apply solely to the offender: showing remorse, having good character, ability to rehabilitate, having a mental disorder/learning difficulty, and age/lack of maturity.

The study used short check list surveys filled out by Crown Court judges for individual cases. Judges indicated whether certain factors (e.g., use of a weapon, lack of premeditation, whether the offender had previous relevant convictions) were relevant for the case. Although the race/ethnicity of the offender was not included in this survey, the researchers were able to obtain this information from a case identifier that matched other court documents containing this information. 

Three of these factors – remorse, good character, and ability to rehabilitate – are more likely to be listed by judges as being important for Whites than for Blacks in both assault and drug cases. Each of these factors is likely to lead to a lower sentence. Remorse, for example, “is a highly subjective factor. Judges have wide discretion in how it is assessed and in deciding if it is a mitigating factor in any given case” (p. 251). If Black and White offenders were to communicate remorse, good character, or their ability to rehabilitate in different ways, it is likely that the predominantly White judges would be more likely to identify it in White offenders than in Black offenders. It is not surprising, therefore, that good character was cited as a mitigating factor in 16% of drug cases involving Whites, but only 11% for Black offenders. Ability to rehabilitate was identified in 15% of White drug offenders but only 9% of Black drug offenders.

Judges do not simply decide on the final sentence based on “facts”. They decide which sentencing factors are relevant and how the offender should be seen on each dimension. It seems likely that factors such as those that are the focus in this study – remorse, good character, and ability to rehabilitate – will explain some of the disparity in sentencing outcomes that we see for different race/ethnicity groups.

Conclusion: This paper demonstrates that in a jurisdiction in which sentencing guidelines are being used to address concerns about variability in sentencing, certain characteristics of offenders may account for race/ethnic variation in the sentences that are imposed. If these factors are to continue to be used by judges to determine the sentence, it is not clear how these racially correlated disadvantages can be eliminated. If, for example, certain groups (e.g., Blacks) are seen as less likely to benefit from rehabilitation programs that are available than are other groups (e.g., Whites), is it fair to sentence Blacks more harshly because of the failure to provide programs that are seen as equally useful for Black and White offenders? It would appear that a more effective approach would be to examine more carefully the factors that judges should be considering when handing down sentences and to ensure that these factors do not, automatically, disadvantage certain groups.

Reference: Guilfoyle, Eoin and Jose Pina-Sánchez (2025). Racially Determined Case Characteristics: Exploring Disparities in the Use of Sentencing Factors in England & Wales. The British Journal of Criminology, 65, 21-260.

Item 6

False confessions made by suspects to the police are not uncommon in criminal matters. Some of the factors leading to false confessions – and false guilty pleas – are known. But in addition, the presence of a false confession can bias other evidence in a case. Enough is known now that we should be able to reduce the likelihood of police-induced false confessions. 

Many of the proven exonerations of those found guilty of criminal charges involve false confessions. A 1992 study found that nearly 30% of the DNA exonerations in the US originally involved false confessions (full or partial admissions of guilt) made by suspects. False confessions leading to inappropriate findings of guilt have been documented in many countries including Canada. This paper is an update of an earlier review of this issue (Criminological Highlights, 11(3)#4). 

“The frailty of human memory poses a significant challenge for investigative interviewers” (p. 23). Most suspects waive their right to silence (see Criminological Highlights 22(2)#8). There are many different types of false confessions including those where the suspect acquiesces to the demand for a confession to escape a stressful situation, or a harsher punishment, etc. In other situations, people under intensive suggestive questioning by the police appear to “develop such a profound distrust of their own memory that they become vulnerable to influence from external sources” (p. 11). Interrogations that are aimed at extracting confessions often start with the objective of increasing the suspect’s anxiety associated with denial and reducing the suspect’s anxiety by encouraging a confession. Interviews often last 6-12 hours with the suspect isolated from others. Often these interviews link the suspect with assertions made with certainty by the police that are linked to evidence – real or manufactured – supporting the suspect’s guilt. In some countries (e.g., the US), courts have ruled that it is permissible for the police to lie to suspects about the evidence. To combat some of these problems, new science-based approaches to interrogation have been developed in recent years. For example, “information- gathering approaches” (p. 22) have been shown to be more effective at determining actual guilt than are accusatorial approaches. 

It is also known that other errors about evidence are especially likely to occur in cases involving false confessions. These include invalid or improper use of forensic science, eyewitness identification errors and informants who have lied (p. 25). In other words, false confessions lead to distortions in other evidence that implicate the suspect. 

This paper proposes a number of plausible remedies that can be used to reduce the likelihood of false confessions or to make it more likely that the confession is identified as being false. These include obvious approaches such as video recording all suspect interviews/ interrogations and ensuring that interrogations are based on solid evidence not hunches. In addition, the situations that are associated with false confessions – such as long periods of isolating the suspect, using tactics that imply leniency if the suspect were to confess – need to be eliminated. Similarly, it is important that confessions not be allowed to taint other evidence. Confessions, therefore, should not be disclosed to other lay or professional witnesses or those investigating the crime. 

Conclusion: Much more is known now than was known 15 years ago when some of these same authors reviewed the ways in which false confessions are induced and the manner in which they lead to false convictions. The remedies that are suggested in this paper are unlikely to eliminate all false confessions. None of them, however, would appear to interfere with the sensible collection of accurate information about the crime that is being investigated.

Reference: Kassin, Saul M., H.M.D. Cleary, G.H. Gudjonsson, R.A. Leo, C.A. Meissner, A.D. Redlich, and K.D. Scherr (2025). Police- Induced Confessions, 2.0: Risk Factors and Recommendations. Law & Human Behavior, 49, 7-53. 

Item 7

In 2012, laws requiring life without parole for youths convicted of murder in the US were found by the US Supreme Court to be unconstitutional. The required release of hundreds of these youths demonstrates, once again, the low rate of recidivism of people convicted of murder who are allowed back into the community.

Previous research has demonstrated that the recidivism of people convicted of murder who are released back into the community is very low (e.g., Criminological Highlights 15(1)#2, 18(5)#8). This paper examines the recidivism rates of youths who were sentenced to life without parole who, because of a court ruling, were eventually released back into the community.

There are a number of reasons to suspect that the committing of a serious violent offence by youths may reflect “a lack of maturity and an underdeveloped sense of responsibility” (p. 2) in most youths rather than characteristics of the youth that would be stable across the life course. The 2012 US Supreme Court decision allowed for this to be tested by examining the behaviour of the 287 youths originally sentenced to life without parole in Pennsylvania who were released back into the community.

Not surprisingly, most (97%) were male. Although Blacks constitute only about 12% of the population of Pennsylvania (and 42% of the population of its largest city, Philadelphia), the vast majority (76%) of those youths sentenced to life without parole who were made eligible for release and who were subsequently released were Black. 54% had been sentenced for first degree murder.  This paper examines the involvement of these 287 youths in offending after release. At the time the data were collected, most (84%) had been in the community for 3 years or more. 15 of the 287 former prisoners (5%) were charged with criminal offences.

There was one murder charge laid against one former prisoner. Five other former prisoners were charged with other person offences (4 aggravated assaults, and one “disarming a police officer”). The vast majority of those who had been sentenced as youths to spend their full lives in prison after a conviction for murder did not, obviously, commit any offences when they were released. This is, of course, not very surprising for two reasons. First, as noted above, it is well established that few of those who commit a murder early in their lives will commit another if they are released back into the community.  Second, by the time that they were released back into the community, they were past the most vulnerable period in their lives for committing violent offences. In this case, only 5% were charged with committing any criminal offence within an average of 4.8 years after having been released. Two thirds were charged only with non-violent offences.

Conclusion: The idea that people who have committed very serious crimes – e.g., murder – must be incarcerated forever to protect society is obviously challenged by these findings. Most of those who have committed a murder either as a youth (this study) or as an adult (other Criminological Highlights papers cited above) and are then released back into the community are likely to live out their lives in a peaceful manner. Although the exceptional cases – e.g., the one person charged with a new homicide offence in this study – are likely to receive a fair amount of publicity, we need to think about this one case in the context of the other 286 who did not commit murder.

Reference: Sbeglia, Colleen and 6 others (2025). Life after life: Recidivism among Individuals Formerly Sentenced to Mandatory Juvenile Life Without Parole.  Journal of Research in Adolescence, 35, 1-11.

Item 8

There is strong evidence that being exposed to high levels of violence in their communities reduces the likelihood that children will experience upward mobility. This paper demonstrates that the reverse may also be true: that depressed levels of upward economic mobility in a community increase the likelihood that residents will be violent.

Previous research has demonstrated that high rates of violence in a community reduce the likelihood of upward economic mobility of its residents. This paper examines the reverse relationship: that depressed economic mobility in a community leads to increases in rates of violence.

“Ethnographic research suggests that community-level violence flows not only from the need to find alternative forms of economic advancement or self-defense, but from a communal milieu of frustration and hopelessness created by blocked opportunity, a structural condition that pushes some community members toward violence” (p. 1).

Based on income tax data from 2,141 counties in the United States, the researchers estimated “the county’s mean adult income rank at age 26 for those born between 1980 and 1986 whose parents were at the 25^th percentile in the national distribution between 1996 and 2000…. Conceptually, this [independent] variable describes the average level of economic mobility… in early adulthood for children who were raised in low-income families…” (p. 7).  The dependent variables – violence and homicide rates – were from county-level Uniform Crime Reports for 2008.

To demonstrate a relationship between depressed mobility and crime, it is important to control for other community variables known to relate to violence levels. Therefore, in this study poverty (as a measure of absolute economic disadvantage), economic inequality (the GINI index, as a measure of relative inequality within the county), the number of law enforcement officers per 100,000 residents, and the unemployment rate (as an indicator of the county macroeconomic conditions) were included as control conditions. In addition, control measures were included that measured outmigration from the community, housing stability, past levels of violence (10-12 years prior to the dependent variable measure) and the state in which the county was located. 

As it turns out, including these control factors made little difference to the overall finding: the measure of intergenerational mobility is significantly related to violent crime and to homicide rates. People growing up in a community where upward mobility from poverty is not likely to occur are more likely to be involved in violent crime and homicide. “While these models do not… identify a causal relationship, they offer initial evidence that [the] results are not merely capturing the reverse documented relationship between violence and later mobility” (p. 5).

Conclusion: The results “offer strong descriptive support” for the hypothesis “that lower levels of intergenerational mobility in communities sparks higher levels of violent crime” (p. 6). Furthermore, it would appear that the size of the reported effects is quite large. The violent crime rate for low-income people who were near the bottom of the mobility measure was roughly twice that of low-income people who grew up in counties with an average rate of mobility. 

Reference: Mann, Olivia, Kathryn J. Edin, and H. Luke Shaefer (2024). Understanding the Relationship Between Intergenerational Mobility and Community Violence. Proceedings of the National Academy of Sciences, 121 (33), 1-8.

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