Criminological Highlights Vol. 20, No. 3 - October 2022

Imprisonment is expensive, both financially and in terms of its effects on prisoners and their families.  In 2018-19, Canada spent an average of $120,589 per federal prisoner per year on imprisonment. Provincial/territorial imprisonment costs averaged $69,350 per year.  

A portion of those who are found guilty could be punished either through a prison sentence or a community punishment.  Is there a reason related to public safety to imprison someone who could be adequately punished in the community?  This review looks at those studies that provide meaningful comparisons on reoffending of those imprisoned compared to similar people who were punished with community sanctions.  A major justification for imprisonment over community punishment is that imprisonment will act as a more effective deterrent for those subjected to it, though the evidence from individual studies does not support this conjecture.  

There are a few experimental studies of the effects of imprisonment (see, e.g., Criminological Highlights 11(4)#2, 3(4)#4). Most studies, however, take advantage of the fact that there is a good deal of overlap between the most serious cases that receive community sanctions and the least serious cases that result in imprisonment.  This paper carefully examined the results of 116 separate studies, representing 4.5 million offenders sentenced in 15 different countries.  For the most part, these studies examined the effect of imprisonment on new convictions, arrests, charges, or reincarcerations.  A small proportion of the comparisons involved other measures (e.g., self-report offending).

Overall, there was a very small effect such that those who were imprisoned were more likely to reoffend than were people who were given community sanctions.  When relevant controls were included, there was still a small criminogenic effect.  Furthermore, the study found “that imprisonment has a null or slight criminogenic effect on reoffending regardless of variations in methodological quality [of the study], the sanctions evaluated, and sociodemographic characteristics of the samples” (p. 387).   However, the small criminogenic effect was reduced or eliminated when full controls were implemented.  Nevertheless, there was no indication of any set of conditions in which imprisonment reduced reoffending. “The minimal effects of custody on reoffending are the same for both males and females and both adults and juveniles” (p. 398).  At the same time, the study does not challenge the view that “Imprisonment can be justified on the grounds of just deserts and incapacitation, but…. the null effect for custodial sanctions [on reoffending] undermines any justification based on specific deterrence” (p. 402).

Conclusion.  “Every review [of the effects of imprisonment on reoffending] has reached nearly the same conclusion: compared with noncustodial sanctions, custodial sanctions, including imprisonment, have no appreciable effect on reducing reoffending….  Based on past research and the findings of this meta-analysis, the limited effects of custodial sanctions on reoffending should be viewed as a criminological fact.  The null effects finding has been replicated repeatedly and independently” (p. 401, emphasis in the original). “Although limited, [other] research shows that harsh or painful prison conditions are not associated with reductions in reoffending and, if anything, are criminogenic….  The core engine underlying effective deterrence is the certainty and not the severity of punishment…” (p. 402).

Reference: Petrich, Damon M., Travis C. Pratt, Cheryl Lero Jonson, and Francis T. Cullen (2021).  Custodial Sanctions and Reoffending: A Meta-Analytic Review. Crime and Justice, 50, 353-424.

Assessment scales in criminal justice are typically evaluated by examining their predictive value in determining whether a person will commit some kind of offence. A typical index is AUC which can be interpreted as “the probability that a randomly selected individual who reoffended received a higher risk score than a randomly selected individual who did not reoffend” (p. 217).  This paper investigates the cross-cultural fairness of an actuarial risk scale, the “Level of Service/Risk, Need, Responsivity” scale (LS/RNR), though the lessons learned go far beyond that specific scale.

There is often concern about whether scales are equally useful for different groups. The problem is more complex than simply determining whether the AUC (or any standard measures of accuracy) is the same for two groups. The “overall accuracy” scores can be the same, but when average scale scores vary and/or base rates of offending vary, a score of a person from one group may mean something very different from the same score of a person from another group.  Simply put, the blind use of these actuarial measures across groups (as if group membership were irrelevant) can easily lead to the over-criminalization of already disadvantaged groups (often racialized groups).

This study examined the reoffending rate (measured by police charges) of 380 serious prisoners in Victoria, Australia who were assessed using the LS/RNR. Almost half identified as Aboriginal and/or Torres Strait Islanders (hereafter referred to as Indigenous).

The LS/RNR scale predicted reoffending for both groups.  The accuracy of its ability to predict being charged by the police for Indigenous and non-Indigenous prisoners was essentially the same (AUCs of 0.60  and 0.63, respectively, which was described (p. 224) as being relatively poor).  But the real problem was hidden from this first analysis.  Further analyses demonstrated that the scale could not discriminate Indigenous non-reoffenders from non-Indigenous re-offenders.  It did better on discriminating non-Indigenous non-reoffenders from Indigenous re-offenders.

Using the scale value alone, then, an Indigenous person who would not, in fact, reoffend, was more likely to be misclassified as someone who would reoffend than would be the case for a non-Indigenous person who wouldn’t reoffend.

Conclusion: Classifying Indigenous people as higher risk when they do not reoffend obviously can lead to harsher monitoring and surveillance. If decisions are made using risk scores or classifications, Indigenous non-offenders would be likely to be treated the same as or as higher risk than non-Indigenous reoffenders.  A starting point for assessment measures, then, might be to develop different norms for different groups.   But more generally, the results suggest that equal “accuracy” rates for assessment scales used by correctional authorities are not sufficient indicators that the instrument is cross-culturally fair.   Although this study focused on one commonly used scale (the LS/RNR), there is no reason to believe that the lesson learned from this Australian study is not important for other classification measures, such as those used for determining the security classification prisoners receive when they enter penitentiary.  Furthermore, the lessons are relevant for other groups (e.g., women and various racialized groups).  Simply put, when groups differ in their prediction scores and/or their ‘outcome’ measures, treating all groups ‘equally’ is a formula designed to ensure unfair outcomes.

Reference: Ashford, Linda J., Benjamin L. Spivak, James R.P. Ogloff, and Stephane M. Shepherd (2022). The Cross-Cultural Fairness of the LS/RNR: An Australian Analysis. Law and Human Behavior, 46(3), 214-226.

There is a fair amount of information suggesting that criminal records that may be many years old can disadvantage people in a manner that serves no social purpose (see Criminological Highlights collection Evaluating the Benefits of Pardons). 

Some jurisdictions such as California have begun automatically suppressing information about certain criminal records.  This can be an effective, and efficient, way of relieving people from the effects of records since it does not require applications, careful assessment, and individual judgements. Broadly speaking, the evidence is clear that after a certain amount of time has passed without reoffending, a person who once offended is no more likely to commit an offence than are those without criminal records.  However, jurisdictions that have automatic systems of relief from criminal records also tend to apply these systems rather selectively.  This study investigates whether the restrictions on automatic criminal record suppression in California end up excluding Black Americans even though they do not explicitly mention race.

One problem with most “application” based systems of relief (where the ultimate decision to offer relief is discretionary) is that they often involve the expenditure of resources (time and money) on the part of the applicant.  These may favour certain groups. Hence “criminal record relief laws may inadvertently provide the greatest benefit for people least impacted by historically punitive policies” (p. 404). California’s automatic relief laws applied to those who had been crime free for a certain period of time after involvement in relatively minor charges/convictions where the defendant had not been sentenced to prison. Arrests without convictions were also eligible for automatic relief.  Most felony convictions were not eligible, nor were cases in which a person violated a condition of probation. If a person received a suspended prison sentence, they were not eligible for automatic relief.  Being eligible for record relief, therefore, is often the result of discretionary criminal justice decisions -- e.g., whether the case is fully prosecuted, whether to seek a misdemeanor or felony conviction, whether to treat a breach of probation formally or not. Hence it should not be surprising to find that Blacks were often less eligible for automatic relief from their criminal records.

For example, Blacks were more likely to have been convicted, rather than having charges dropped. This one factor excluded a higher proportion of Blacks than Whites from automatic relief.  Blacks were more likely to have felony convictions in their record and they were much more likely to have been imprisoned.  These factors and the restrictions on automatic record relief mean that a lower proportion of Black Californians with convictions are eligible for an automatic record suppression. It was found that an automatic “7 year sunset rule” (where the record would automatically disappear after 7 years) would have the greatest impact on Black Californians.  Though in general White Californians would benefit more from automatic relief laws, granting automatic relief for those offences where relief is only available on a discretionary basis and providing automatic relief for those who were 7 years past their convictions would reduce the Black-White difference in eligibility for relief substantially.

Conclusion:  This paper provides a criminal justice illustration of how apparently racially neutral policies (offence-based criteria on eligibility for criminal record relief) may disproportionately help certain groups (in this case, White Californians).  Relatively small amendments to these provisions (e.g, broadening the scope of eligible offences, or implementing relief automatically after a specified period of time) can reduce the size of the disadvantage to certain groups.  But more generally the paper suggests that it is useful to examine the effects of justice policies on different groups before implementing these policies.  “Considering the racial impact in the construction of a policy will make it easier to confront and avoid harm” (p. 415).

Reference: Mooney, Alyssa C., Alissa Skog, and Amy E Lerman (2022).  Racial Equity in Eligibility for a Clean Slate Under Automatic Criminal Record Relief Laws.  Law & Society Review, 56, 398-417.

Although it is relatively well established that solitary confinement has adverse effects on the mental health of prisoners (Criminological Highlights 19(6)#4), there is less information about the impact of the mental health of a prisoner on the likelihood that they will spend time in solitary confinement.  It may be that the causal arrow points in both directions.

This paper looks carefully at the impact of mental illness on time spent in solitary confinement.  Canadian research suggests that those with deteriorating mental health are especially likely to spend long periods of time in Canada’s recently renamed solitary confinement cells.  Solitary confinement, it seems, is used as a means of penal control and as a way of managing the prison population.

This study, which looked at the prison records of over 90,000 Pennsylvania prisoners, focused on the mental state of prisoners at intake to prison.  Prisoners were assessed as having one of four degrees of mental illness:  none, prior but not current mental health diagnosis, current mental illness diagnosis, current treatment for serious mental illness.

The goal of the study was to look at the relationship of the prisoner’s mental health status at the start of the sentence to the likelihood that they would be forced to experience solitary confinement.  Pennsylvania officially uses solitary confinement for punitive purposes (e.g., misconducts) and for administrative purposes (for those who are described as threatening orderly prison operations).  About 22% of prisoners spend as least some time during their sentence in punitive segregation and 26% spend some time in administrative segregation.

For both men and women, the more serious the prisoner’s mental health problems were at admission, the more likely they were labelled as exhibiting prison misconduct.  Furthermore, as the mental health status moved up the scale (from none to serious mental illness), the likelihood of disciplinary custody during the sentence increased dramatically (from 17% to 31% for men; 9% to 23% for women) as did the likelihood of administrative custody (from 22% to 41% for men; 9% to 17% for women).  Not only were there more misconducts registered against those with mental illness, but the likelihood of disciplinary custody, given a misconduct, also increased substantially.

When various controls were introduced – e.g., demographic variables, risk scores, offence severity, misconduct severity -- the presence of mental illness (especially serious mental illness) still increased the likelihood of being placed in solitary confinement for both men and women.  “Disproportionate solitary confinement results mostly from the large number of prison misconducts tickets [records of misconduct] written by prison staff to mentally ill prisoners. Tickets are mostly written for the nonviolent misconduct of threats and defiance…. These results imply that common markers of criminal justice inequalities, such as disparities in overall incarceration rates, underestimate the burden of harsh prison conditions for people with serious mental health problems” (p. 559).

Conclusion: It is clear that the causal arrow between experiencing mental illness and experiencing solitary confinement points both ways.  Defiant or threatening behaviour – which appears to lead to solitary confinement -- “could be artifacts of mental illness” (p. 560).  These “findings of mental health disparities in solitary confinement show how American prisons heap the harshest forms of punishment on the most vulnerable” (p. 561).

Reference: Simes, Jessica T., Bruce Western, and Angela Hill (2022).  Mental Health Disparities in Solitary Confinement.  Criminology, 60, 538-575.

“Racial disparities have been documented at virtually every stage of criminal justice processing, from investigation to sentencing” (p. 783).  This paper looks at the way in which professional recommendations about whether a prisoner should be released on parole can create racial disparities in release decisions.

Due process protections and rules of evidence are minimal in parole decisions.  Issues such as who the prisoner speaks with while in prison, their mental health history, etc., can be brought to the decision maker’s attention directly and indirectly.  The theory is that this “wealth of information” will allow “decision-makers to develop holistic understandings of parole candidates as individuals” (p. 785). But “holistic understandings” are often large enough to hide information that creates racial disparities in outcomes.

In California parole hearings, decision makers have access to a breadth of information (about the applicant’s background, medical records, offence and trial and sentencing information, employment history, victims’ views, etc.), as well as detailed information about their behaviour in prison (e.g., rehabilitative efforts made by the prisoner that might be recorded by prison authorities).  “Parole boards… have access to more information than sentencing judges do” (p. 789).   But in addition, parole authorities in California have access to psychologists’ assessments of dangerousness, disciplinary infractions identified and recorded by correctional personnel, and prosecutors’ reports of their opposition to parole.  In other research, psychologists’ assessments have been shown to advantage White defendants and to disadvantage Blacks.  The other two “expert” sources may also disadvantage Black applicants for parole.

Transcripts of 680 parole hearings for men applying for parole from life sentences (mostly for murder) were coded.  White applicants were more likely to be successful than Latinos and Blacks (17% vs. 14% and 9%, respectively). Prosecutors were considerably more likely to oppose parole for Black applicants than for the members of the other two groups.  Similarly, Black applicants were less likely to be rated as low risk and were more likely to have disciplinary infractions on their records.

Controlling for various factors (e.g., age, criminal convictions, number of victims, whether they earned a degree while in prison) Black prisoners were still less likely to be granted parole.  The racial differences became statistically non-significant, however, when the assessments of the experts (psychologists, those recording disciplinary issues, and prosecutors) were controlled for.  In other words, it would appear that the different success rate at parole by the three groups may have been a function of the recommendations by these three groups of ‘experts’.  Other analyses suggest that “the disparity [in outcome] does not seem to be related to racial differences in self-rehabilitation” (p. 807) – getting a degree or finding a job.  It appears that the parole commissioners relied on “prior professional evaluation”, suggesting that the Black-White disparities “are not due to differences in the rate at which lifers pursue rehabilitation…” (p.  807).

Conclusion: Black applicants for parole may not only be seen as more “criminal” than White applicants, but they may also be seen as less truly rehabilitated by those running the prisons, psychologists who assess them, and prosecutors.  “Nearly half of the racial inequality [of parole application outcomes] is associated with parole commissioners’ reliance on prior professional assessments” (p. 808).  The problem is that “highly subjective judgements, such as psychiatric reports and disciplinary infractions, become part of the official record…. Every file contains exercises of discretion [that are described] as ‘facts.’  If a prison guard believes someone is disrespectful and gives him a citation, that parole candidate ‘has’ [a citation]” (p. 809). This “fact” then affects the prisoner’s freedom.

Reference: Young, Kathryne M. and Jessica Pearlman (2022).  Racial Disparities in Lifer Parole Outcomes: The Hidden Role of Professional Evaluations.  Law & Social Inquiry, 47(3), 783-820.

When a person is arrested, they are often required to appear in a court in order to be released.  Whether the legal question is whether the person should be released, or how much money needs to be posted in order for them to be released, their friends and family – who may need to attend this court for the accused to have a chance at being released – begin a process whereby they are punished. In other words, the breadth of criminal justice punishment extends beyond the person actually charged.

Previous research (see Criminological Highlights 17(6)#1) has demonstrated that accused people are punished in a number of important ways in addition to any formal sentence they may, in the end, receive.  This paper looks at the experiences of friends and family of an accused in trying to help the accused person be released.   In this study of the “bond court” in Chicago, the punishments experienced by an accused person’s “supporter” may not be intentional, but they are real.  Obviously one way in which family and friends of an accused are punished is through the extraction by the court of the funds that need to be deposited in order for the accused to be released.  Similarly, there are legal costs – often borne by a supporter of the accused – related to mounting a defence.  In addition, accused persons often have to appear in court regularly until their case is disposed of.  Often such appearances involve supporters who not only offer moral support to the accused but may be necessary in providing transportation to the court.

Criminal courts – perhaps most notably bail or bond courts – operate in a manner demonstrating that they do not give any value to the time of an accused person or their supporters.  People may wait in court all day for a hearing that may last only a few minutes.  In addition, often supporters experience embarrassment at being associated with a person accused of a crime, as well as degradation by court officials in part because they do not necessarily know “the rules” by which courts operate.

In this study, courts were systematically observed and supporters of the accused were interviewed.   If supporters needed to post money to achieve release, this could have a dramatic impact on them if they were poor, especially since the money would be tied up for months or years. Transportation costs to get to the court may be non-trivial.  And, in the case of the central Chicago court, parking costs were not only high, but limited to a period of 2 hours. If a case was not completed within 2 hours of the supporter’s arrival, the parking fee had to be renewed, requiring the supporter to leave the court to renew the parking fee.  Because of the uncertainty of when a case might be called, supporters often faced expensive parking tickets.  Even though mobile phone payment of parking was available, this was not available to supporters of the accused because cell phones were not allowed in court.  This restriction also meant that those who might be able to be productive using silent features of their cell phones were not able to do so. Supporters who had jobs often lost income – often a full day’s income – because of the unpredictability of when a case would be heard.  Supporters suffered because of the unpredictability of the court (it seldom started on time) and the lack of information about how things work. Often it was impossible for supporters to hear or understand court proceedings.  

Conclusion:  Courts extracted resources from supporters of an accused (explicitly or as unintended consequences). Courts ordered changes to living arrangements (requiring an accused to live with or avoid certain people) and frequently created fear and shame, in effect “punishing [supporters] by degrading them and their relationships with people charged with crime.  The experience of attending court as a supporter “should be understood as being punitive” (p. 603) even though the actions of the court are not normally described as punishments. Most of these effects disproportionately punished the poor.

Reference: Eife, Erin and Richie, Beth E. (2022).  Punishment by Association: The Burden of Attending Court for Legal Bystanders. Law & Social Inquiry, 47(2), 584-606.

In Canada in 2019, there were 39,341 cases in which a person failed to appear, as required, in a criminal court.  In most (83%) of these incidents, the person was charged with the criminal offence of “failure to appear” (FTA).  This paper, building on another study in which people were sent post-card reminders of their required appearance (Criminological Highlights 13(4)#1), examines the effect of a simple telephone reminder on the rate that people show up for required court appearances.

The failure of an accused person to appear, as required, in court is expensive in a number of different ways.   First, time in criminal court is very expensive given the number of people who are necessary to run a criminal court.  Second, a failure to appear in court can land the accused in custody awaiting trial.  This has obvious costs both to the accused person but also to the community given the high cost of incarcerating someone.  Third, a conviction for FTA can increase the likelihood of pretrial detention in the future and can lead to harsher sentences in the future.

In New York City, after the bankruptcy of a private vendor who had the responsibility to remind people of their court dates, the city found that it could not, immediately, remind everyone of their court dates.  Instead, it set up a study where people were randomly assigned to one of four conditions:  (1) No reminder.  (2) A reminder telephone call three days before the scheduled court date. An attempt was made to speak to someone, but if, on the second attempt, this was not accomplished, a voice message was left, if that was possible.  (3) A call the same day as the required court appearance.  Calling began at 6 am and ceased at 9 am.  Court started at 9:30 am.   (4) Both a 3-days in advance call and a same-day call.  The caller was reasonably successful (62% to 76%) in at least leaving a message for the accused person.  The main data analysis looks at the original (randomized) assignment to condition, regardless of whether the attempt was successful.

In the control group, 19.3% of the accused people failed to appear for their scheduled court appearance.  The three groups that received reminders did not differ in their appearance rates. Their average failure rate was 12.1%, a 37% reduction in the rate of FTAs.  The effect of the reminders was greater for Black, Hispanic, and Asian accused people than it was for White accused people.  Furthermore, the effect of the reminders was greatest for those facing multiple charges.  For multiple-charge cases, the FTA rate for those who did not get reminders was 22.4%.  Reminders reduced this to 11.4% (a 49% reduction in the FTA rate).  Finally, the effect was largest for those who had long wait times between their initial police contact and their appearance in court.

Obviously, it was necessary to get a phone number for each accused person.  But the cost of the telephone calls was not high.  In this study, attempts were made to contact 1202 people.  Attributing all of the costs of the calls to the cases in which the accused showed up but, without the call, probably wouldn’t, it was estimated that each FTA that was avoided cost the city about $34.  Unfortunately, no data were provided on the court costs associated with a FTA, but rough estimates would suggest that Ontario courts cost at least $20 per minute to run.   But in addition, the policing, detention, and additional court appearances required to process a new charge suggest that a simple phone call pays for itself. These results are similar to findings showing that a virtually cost-free re-design of court appearance documents can reduce FTAs (Criminological Highlights 19(1)#3).

Conclusion: The data are clear: Telephoning accused people prior to their court appearances can reduce dramatically the number and costs associated with failures to appear.  But in addition, the benefits are greatest for members of racialized groups and those facing larger numbers of charges.

Reference: Ferri, Russell (2022).  The Benefits of Live Court Date Reminder Phone Calls During Pretrial Case Processing.  Journal of Experimental Criminology, 18, 149-169.