Criminological Highlights Vol. 15, No. 5 – March 2016

View this issue as a PDF: PDF iconCrimHighlightsV15N5.pdf

This issue of Criminological Highlights addresses the following questions:

  1. Where do gangs get their guns?
  2. What kinds of strategies can the police use to reduce crime in specific neighbourhoods?
  3. Can employment programs for those being released from prisons reduce reoffending?
  4. Who has a bigger effect on criminal sentences: the judge or the prosecutor?
  5. Can we assume correctional programs that sound as if they should be effective will do no harm?
  6. How do risk assessments affect the pretrial release process?
  7. Did the softening of drug laws in Portugal change dramatically drug use in that country?
  8. How do different types of conservatives differ in their views of the police?

Guns used by gangs in Chicago are largely imported from other states and tend to have had a number of previous owners.

Cities with high homicide rates tend to focus their attention on trying to control access to firearms since a large proportion of homicides, in many locations, are carried out with firearms. In the US, 68% of homicides were carried out with firearms. In Chicago, where this study took place, about 80% of homicides involve firearms.  The comparable figure for Canada is 31%. 

Controlling gun availability is often seen as an effective – or perhaps the major – method of lowering the rate of firearms injuries and deaths.  However, to control availability, one needs to understand where the firearms involved in injuries and deaths come from.   This study uses information about the original purchasers of firearms confiscated by the police from Chicago gang members.  Guns on the street, in Chicago, are apparently much more expensive to obtain than the same gun from gun dealers, reflecting, presumably, the relative difficulty in obtaining guns illegally.  New guns in the US are almost exclusively sold by licensed firearms dealers who typically need to be able to identify the purchaser and determine that the person is not disqualified from owning the firearm.  However, private transactions after that initial purchase are less well regulated, or are not regulated at all.  A survey of US prison inmates, for example, suggested that the main sources of firearms (73%) for them were friends, family, or purchases on the street none of which involved questions about the qualifications of the purchaser.   Few firearms came from retail stories (12%) or other sources (14%).  

When a firearm was confiscated by the Chicago police between 2009-2013, a request was made to a federal agency for a trace of its origins.  64% of the 8410 requests were successful.  Although the gang members themselves were young, the guns they were caught with were old – an average 11.6 years had elapsed between the initial purchase and the confiscation. Only 10% of the guns confiscated from gang members were less than 2 years old.  Very few (7.8%) of those who had possession of the firearm were the original (registered) owner. Many (about 20% of those where the gender of the purchaser could be identified) of the guns confiscated from male gang members were purchased by women. 

Illinois has, for many decades, required a background check on purchasers of firearms from gun dealers.  This was not the case for many states, particularly those in the deep south.  In 1994, however, all US gun dealers were required to do a background check. Prior to this change, it seemed that many (35%) of the guns recovered in Chicago came from the deep south, presumably because of the ease in purchasing them there. After gun dealers in the deep south were required to examine purchasers, the proportion of Chicago guns recovered by the police that originated in the south dropped to 15%.  Gun dealers themselves, through direct sales to gang members, do not appear to be an important source of gang guns:  “Direct, well-documented sales of guns by dealers to gang members account for less than 2% of the total [of guns recovered by the police]” (p. 723). “Most guns taken from gang members in Chicago pass through the hands of at least one intermediary – a third party that helped the gun move from dealer to gang member” (p. 725). Indeed, “trafficking is more common for guns that wind up confiscated from gang members than non-gang members” (p. 753). 

Conclusion: It may be that gang members who are responsible for shootings in Chicago have relatively easy access to guns.  But these guns have been around for more than 10 years after they were first sold. Even though there are widely reported preferences for new guns (and guns on the street seem to cost more than guns bought from dealers), the fact that ‘gang guns’ tend to be old suggests that the underground market in guns is not as efficient as the legal market.  Said differently, in a location where there is relatively little control on firearms, regulation and enforcement does seem to restrict somewhat the access to the most preferred guns. 

Reference: Cook, Philip J., Richard J. Harris, Jens Ludwig, and Harold A. Pollack (2014). Some Sources of Crime Guns in Chicago: Dirty Dealers, Straw Purchasers, and Traffickers. Journal of Criminal Law & Criminology, 104 (4), 717-760.

Back to top

Policing strategies that focus on local forms of disorder can be reduce crime. However, aggressive order maintenance strategies that target individual disorderly behaviours appear to be ineffective.

Dealing with physical and social disorder appears to be a central feature of some police services’ crime prevention strategies.  Whether such strategies are effective, however, is contentious (Criminological Highlights 1(4)#5, 4(5)#4, 5(1)#6. 8(4)#1, 8(5)#8, 9(1)#2, 10(3)#4, 14(5)#3), in part because the strategies used and the problems that are targeted vary considerably. 

In this paper, a very thorough search of the published and unpublished research literature on the policing of disorder took place. All adequately designed studies addressing the question of whether policing disorder reduces crime were examined.  To be included in the review, there had to be some kind of ‘control areas’ within the cities. Hence in all cases, the disorder-oriented police strategy was compared to a control area where, typically, policing took place in its normal fashion.  Twenty-eight studies reporting 30 independent tests of policing disorder interventions were found. Two studies were from the U.K.; the rest were carried out in the U.S.  In 9 of the studies, a random, controlled, experimental design was used. In the rest, an attempt was made to find equivalent areas in which the intervention did not take place.

Across all 30 tests, there was a significant effect of police interventions.  Those areas in which social and/or physical disorder was targeted tended to have statistically significantly lower crime rates. However, the effect is described as ‘modest.’  To get an idea of what ‘modest’ means, imagine that in a targeted area, there were, on average 50 criminal incidents a month prior to the intervention being instituted, and that this varied such that for most (95%) of the months we would expect to find  between 40 and 60 incidents.  The targeted police interventions described in these 30 studies would be expected to reduce the number from 50 to approximately 48.85 incidents. 

However, only the community problem solving programs instituted by the police demonstrated significant crime reduction.  Using the hypothetical example above, these programs would be expected to reduce the number of crimes from 50 to 48.6 crimes per month.  

Those programs that attempted to carry out ‘aggressive order maintenance’ programs (e.g., focusing on minor forms of disorder such as public drunkenness, prostitution, vandalism, disorderly youth, or traditional arresting of those thought to be gang members) did not show statistically significant effects.

The effective types of programs seem to have had fairly similar impacts on violent crime, property crime, and disorder and drug offences. 

Conclusion:  Policing that focuses on “community problem-solving that seeks to change social and physical disorder conditions at particular places produces [statistically] significant crime reductions” (p. 581) though these effects are relatively modest in size.  “When considering a policing disorder approach, police departments should adopt a ‘community coproduction model’ rather than drift toward a zero-tolerance policing model, which focuses on a subset of social incivilities….” (p. 581). This latter approach appears to be ineffective.

Referernce: Braga, Anthony A., Brandon C. Welsh, and Cory Schnell (2015). Can Policing Disorder Reduce Crime? A Systematic Review and Meta-analylsis. Journal of Research in Crime and Delinquency, 52(4), 567-588.

Back to top

A comprehensive employment assistance program for prisoners returning to the community reduced re-arrest, but not re-imprisonment.

Most of those who enter prisons eventually are released. Unfortunately, however, even those who are skilled and motivated to find work quickly discover that jobs are particularly hard to get if one has a criminal record (Criminological Highlights 6(3)#2).  These two facts – along with the assumption that former prisoners need to work in order to survive in the community – suggest that it is important to find ways of increasing the likelihood that prisoners will be able to find and hold work when they are released.

Work programs for former prisoners have a mixed history.  Some programs appear to work for certain people (Criminological Highlights 4(3)#6, 3(4)#2, 13(3)#5), but it would seem that simply providing skills to prisoners or finding them an initial job may not be enough.  Recently it appears (see 15(3)#5) that more comprehensive programs that begin while a person is still in prison and continue providing support after the prisoner is released may be more effective.  This paper reports the results of an experiment in which prisoners were randomly assigned to a program with ‘wrap around’ employment services starting in the prison or to receive the services that were normally available for those being released.

The program evaluated in this paper involved 6 months of programming before the prisoner was released. This consisted of both vocational assistance and social-work interventions (3-4 months of programming designed to change risk factors thought to be related to future crime). For example, they had especially easy access to alcohol and drug programs. At the time of release, plans were put in place for those who were randomly assigned to the special comprehensive program to receive assistance in housing, transportation, obtaining relevant documents, and job searches.  Those randomly assigned to this treatment program were given automatic direct access to all these programs that, theoretically, were available to all state prisoners; those assigned to the control condition had (theoretical) access to them, but had to wait until there was a spot.    

The data from both groups demonstrates the challenges facing those trying to help reintegrate prisoners. About a third of each group had less than a high school education. Most owed restitution, and more than half had no funds when released.  Those assigned to the ‘experimental’ high-services group, on release, were eligible for work that, for about 6 months, was subsidized by the program.   Not surprisingly, prisoners in the treatment group were more likely to be employed during the first year.  However, the remarkably low median total income that was earned during the entire first year - $462 for the control group, $2690 for the treatment group – underlines the challenge facing those concerned with successful reintegration. Even those receiving special attention did not, on average, earn enough to subsist in society.   By the end of the first year, 72% of the control group and 63% of the treatment group had been re-arrested (a significant difference).  There was not a significant difference in the likelihood of re-incarceration. 

Conclusion: There is some support for the conclusion that wrap-around services – including employment services – can reduce recidivism.  From previous research it appears that “offering low-wage menial jobs and job-placement assistance without much help in other domains is not promising for achieving the goals of recidivism reduction or sustained increase in employment” (p. 377).  The problem may be that “released prisoners who have never had a regular connection to the labour market may be largely indifferent to [an offer of a temporary low-wage menial job offer], and [the] results on earnings seem to confirm that view as the norm…. A stronger test [of the value of an employment program for prisoners] would require a bigger ‘dose’ of legitimate opportunity [in the area of employment]” (p. 378).

Reference: Cook, Philip J., Songman Kang, Anthony A. Braga, Jens Ludwig, and Mallory E. O’Brien (2015).  An Experimental Evaluation of a Comprehensive Employment-Oriented Prisoner Re-entry Program. Journal of Quantitative Criminology, 31, 335-382. 

Back to top

Even in highly structured sentencing guideline systems, the prosecutor in charge of a case has a large impact on the sentence that an accused receives.

In many jurisdictions, prosecutors and defence lawyers agree on a “joint submission” to the judge on what the sentence for an accused should be.  However, guideline systems often restrict the range of possible sentences that the parties can recommend just as mandatory minimum sentences restrict judges.  This study examines the variation that exists in the sentences handed down in a legal setting (the US federal guidelines) where the most obvious power of the prosecutor is to tell the judge that the person found guilty gave help to the prosecutor.

Federal prosecutors in the US can recommend a downward departure from the guideline if they assert that the accused has given ‘substantial assistance’ to the prosecutor (e.g., by naming the source of drugs that were found on the offender). This study was carried out using data from 1998-2000, when the US federal guidelines were seen as mandatory rather than advisory (as they have been since 2005). The paper examines the influence of the judge and the prosecutor on the sentence that is imposed.

The study uses data from 3 locations in which 18 judges sentenced 2686 offenders to prison. There were 97 different prosecutors and 431 judge-prosecutor combinations.  Various factors (e.g., the presumptive sentence for the offence under the guidelines, criminal record, race, type of offence, plea) were controlled for.   

Controlling for all available case characteristics, sentences varied depending on the judge and the prosecutor assigned to the case.  But above and beyond variation attributable to judges and prosecutors independently, it appears that the combination of judge and prosecutor makes a difference.  In other words, certain combinations of judges and prosecutors led to sentences that were not explainable by looking only at the identity of the judge and the prosecutor.  Particular combinations of judges and prosecutors appeared to lead to harsher or more lenient sentences than one would expect. 

Conclusion: Even at a time when the guidelines were considered mandatory, “Sentences varied by 6.5 months based on the judge [who sentenced the offender] and by more than a year based on the prosecutor…. Sentences also varied based on the judge-prosecutor dyad; in fact… the variation attributable to the judge-prosecutor dyad was greater than the variation attributable to the individual judge or prosecutor” (p. 615).  Given the power of the prosecutor under US guidelines, it is not surprising that “the results reveal greater disparity in sentencing based on the prosecutor than based on the judge.  This finding reflects the fact that the federal sentencing guidelines [prior to 2005] severely constrain the discretion of judges but do little to restrain the exercise of prosecutorial discretion in making decisions that affect sentence disparity” (p. 616). 

Reference: Kim, Byungbae, Cassia Spohn and E.C. Hedberg (2015). Federal Sentencing as a Complex Collaborative Process: Judges, Prosecutors, Judge-Prosecutor Dyads, and Disparity in Sentencing. Criminology 53(4), 597-623.

Back to top

Some criminal justice treatments occasionally cause harm and some treatments that generally are seen as effective have, in many instances, no impact whatsoever.

Most criminal justice interventions are never adequately evaluated.  Since people change over time without formal interventions, adequate evaluations almost invariably require some comparable comparison group that usually receives whatever treatment is normally available to people in that circumstance – the so-called ‘treatment as usual’ comparison group.

In addition to the possibility of a program having no positive effects (and hence to not only be a waste of an opportunity to intervene effectively but also a waste of resources), there are a number of examples of interventions that have been demonstrated to have negative impacts on those who receive them (Criminological Highlights 5(4)#1, 6(2)#4).  This paper examines a number of systematic reviews that have been carried out of criminal justice interventions that have been adequately evaluated to examine whether harmful effects occur – defined in this case as interventions that increase the likelihood of subsequent offending. In this paper, systematic “Campbell Collaboration” reviews where some measure of individual offending was included as the outcome variable were examined to determine just how many programs had been shown to have favourable, null, and unfavourable effects.  These reviews covered 15 different types of interventions such as cognitive behavioural therapy for offenders, mentoring programs,  drug courts, anti-bullying programs, and serious juvenile offender programs.  

One difficulty in drawing strong inferences about what constitutes a ‘good’ or a ‘bad’ intervention is that there are substantial differences across studies within a given ‘type’ of intervention.  For example, drug courts vary in the manner in which they are operated, the type of person who is assigned to the intervention, the type of treatment that may be associated with the court, and, of course, the nature of the comparison group that is used.  A simple ‘box score’ of the outcomes of drug court evaluations, therefore, may not be very helpful in predicting whether a given drug court will have a good, a bad, or no impact on future offending by those subjected to the drug court itself.  To be specific, this review found 71 drug court evaluations showing desirable effects, 78 showing no effects, and 5 demonstrating undesirable impacts. It is likely that there is no simple “drug court” model that has been used; instead, it may be that many of these 154 drug court programs have little in common with one another other than the fact that courts and those who have apparently used illegal drugs are involved.  Hence, it would be necessary, if one is interested in developing useful interventions, to ensure that any new program be properly monitored.  One cannot assume that just because many programs of a certain type have been shown to be effective, that a newly developed program that resembles those other programs will also be effective. 

In this review, most (n=391) of the interventions that were properly evaluated had no positive or negative impact though a fair number (n=232) had statistically significant positive impacts.  Twenty-two, however, had undesirable impacts on those subjected to them.

Conclusion: “Unless social programs are evaluated for potential harm as well as benefit, safety as well as efficacy, the choice of which social programs to use will remain a dangerous guess” (p. 246).  More generally, “High quality designs are needed to be able to identify any harmful or beneficial treatment effects” (p. 246).  Given that programs vary across location and different instances of the same type of program may be offered to different types of people, it is difficult to determine in advance that any specific type of intervention is certain to be ‘good’ or ‘bad.’  However, given the range of outcomes that were found for 15 types of intervention, it is clear that those who rely on testimonials or good motives to justify a program are encouraging society to be wasteful of resources or worse.

Reference: Welsh, Brandon C. and Michael Rocque (2014). When Crime Prevention Harms: A Review of Systematic Review. Journal of Experimental Criminology, 10, 245-266.

Back to top

The use of risk assessment instruments for youths in determining pretrial release has an unanticipated effect: it reduces the use of pretrial detention.

The usefulness of risk assessment instruments in the criminal justice system is usually measured by whether or not they predict future misbehaviour.   This study suggests that they may serve another function: reducing the use of pretrial detention.

Typically, decisions about whether to release someone who has been arrested are based on subjective judgements.  A ‘risk’ analysis of these decisions suggests that decision makers are likely to be ‘risk averse’ and,  as a result, detain people who, in fact, are unlikely to misbehave in any important way (Criminological Highlights 11(1)#6).  An ‘objective’ risk assessment tool, then, may improve the decision-making process by giving comfort to decision makers that a decision to release is justified. 

When making pretrial release decisions, it is often impossible to obtain the full range of information that might be useful for a risk assessment.  The result is that risk assessment instruments (RAI) at this stage of the process are often developed ‘by consensus’ rather than as a result of an empirical investigation. They may, therefore, include factors that do not in fact predict whether or not an accused will appear in court when required or commit an offence while on release.  

In 2004, New Jersey developed a RAI for those making decisions on pretrial release for youths. It included 7 items related to the current charge, criminal record, and previous compliance with court orders. The goal of the study was to see whether the use of the RAI affected the proportion of youths who were detained. 

Cases before and after the introduction of the RAI were matched so that equivalent samples of accused people could be compared on whether or not the accused was released.   The “after” cases came from a period when the decision making tool had become standard practice.  The before and after cases were equivalent, then, on past and current offences, history of noncompliance with court orders, age, gender, race/ethnicity, and the time of day when the decision on detention was made.

Prior to the implementation of the RAI, the detention rate had been 67%.  After the RAI was in place, the detention rate dropped to 40%. The size of this effect could have been influenced by other initiatives implemented at the same time including the provision of new alternatives to detention.  When using the risk assessment tool,  decision makers appeared to give more weight to various ‘objective’ factors in the detention decision, including the number and nature of the charges, and previous failures to appear. Whether the effect was purely the result of the RAI or was due to other policy-based interactions, the results do demonstrate that rates of pretrial detention can be reduced considerably. 

The largest impact appeared to be on the ‘average risk’ juveniles. They were considerably more likely to be released after the RAI was implemented than before, unless they were charged with very serious offences.  There was little impact of the RAI on low risk youths who, in both periods, tended to be released unless they committed a very serious offence.  There was little impact on high risk youths who were typically detained. 

Conclusion: It would appear that providing pretrial release decision makers with an ‘objective’ risk assessment increased the likelihood of pretrial release, especially for those who were of average risk.  The use of risk assessment instruments, then, can be seen as having at least two separate goals: to provide decision makers with an objective assessment of risk but also to reduce the detention of those who are not objectively high risk. 

Reference: Maloney, Carrie and Joel Miller (2015).  The Impact of a Risk Assessment Instrument on Juvenile Detention Decision-making: A Check on “Perceptual Shorthand” and “Going Rates”?  Justice Quarterly, 32(5), 900-927. 

Back to top

Portugal’s decriminalization of possession of small amounts of drugs in 2001 had little impact because most of the changes had occurred prior to the change in the law. 

Previous research (see Criminological Highlights 11(5)#7) has demonstrated that when Portugal decriminalized the possession of small amounts of drugs in 2001 there were few measurable impacts on drug use or drug prosecutions.  This paper examines the reasons that so little change took place notwithstanding the rather dramatic change in the law. 

In 2001, Portugal decriminalized the acquisition, possession and personal use of small quantities of all psychoactive drugs.  One of the purposes of the change was to destigmatize addicted users and encourage them to seek treatment. Other aspects of the drug laws (e.g., production, distribution) did not change, nor did the new law permit or regulate use. In other words, to obtain drugs, one still had to rely on illicit markets and one still risked being apprehended by the police.  Portugal’s more liberal approach to drugs (i.e., decriminalizing simple possession of drugs) is consistent with policies in other European countries (e.g., Spain and Italy) that do not impose criminal sanctions for possession of small amounts of drugs. 

The orientation of Portugal’s criminal law prior to decriminalization “was symbolic rather than punitive… and [was] designed to encourage [the drug consumer] to seek treatment” (p. 751). The new law decriminalized the possession, for purposes of personal use, of an amount of any psychoactive drug deemed to be what a single person would use in 10 days.  However, those possessing an amount below the thresholds for criminal intervention were still issued an administrative citation.

The number of police contacts with drug users (arrests and citations combined) did not change appreciably after the legal change was made, though most of these contacts shifted from resulting in court appearances to appearances before non-criminal bodies. However, the number of drug trafficking convictions declined by 27% and the year-end count of those in prison for drug offences dropped by 47%.  It would seem, then, that the criminal justice system “embraced de facto practices of greater leniency for at least some drug users and purveyors whose behaviour remained criminally sanctioned… . The drug reforms were, in part, an articulation of shifting sentiments and codification of penal practices already occurring with respect to drug use” (p. 758).  

Survey data on recent use of any drug went up very slightly (from 3.4% admitting some recent use in 2001 to 3.7% in 2007).  Use of a psychoactive substance some time in one’s life went up more: from 7.8% to 12%.  The use of cannabis (and other drug use) measured by student surveys increased during the 1990s, but levelled off in 2003. 

Conclusion: “The reforms [in Portugal] were not a move toward liberty [in the use of drugs] but a shift from one arena of government involvement to another” (p. 775).  Other drug-related offences were left unchanged in the law. Nevertheless, there appeared to be fewer drug cases going into court even among those infractions that were still criminalized.  The law largely removed criminal sanctions and substituted an invitation to seek treatment. In effect, then, it would appear that the changes in law reflected changes in Portugal’s culture with respect to psychoactive drugs that had largely already occurred, rather than causing great changes in the manner in which those involved with psychoactive drugs were dealt with.

Reference: Laqueur, Hannah (2015). Uses and Abuses of Drug Decriminalization in Portugal. Law & Social Inquiry, 40(3), 746-781.

Back to top

‘Conflicted conservatives’ – those who identify themselves as being conservative, but who have liberal spending preferences – look like self-identified liberals in their opposition to police use of force.

Conflicted conservatives – people who identify themselves as being political conservatives, but who favour liberal spending policies over conservative policies – comprise about 20% of the American electorate and about half of the self-identified conservatives.  Because they appear less committed to voting for a particular political party, they are, from an electoral perspective, an important group. This paper examines their views of police use of force.

Conflicted conservatives are utilitarian in their outlook and, therefore, are relatively likely to vote for a candidate – of any party - who is likely to deliver the outcome they desire.  Criminologists often distinguish between utilitarian concerns about crime and “symbolic beliefs arising from worldviews, moral outrage or retributive desires” (p. 653).  It would be expected that conflicted conservatives views of criminal justice would be driven by utilitarian concerns rather than retributive desires.

Using US General Social Survey data collected between 1972 and 2014, respondents were identified as being liberals, moderates, conflicted conservatives, or consistent conservatives.  Their place on the overall political spectrum was determined by their indication of where they stood on a 5-point scale ranging from “extremely conservative” to “extremely liberal” (p. 656). Consistent and conflicted conservatives were distinguished according to whether or not they favoured increased spending on issues such as improving and protecting the environment, improving the nation’s education system, and improving the conditions of Blacks.

Support for police use of force was assessed with a series of questions about police use of force.  “Global support” was assessed with a question that asked whether respondents “could imagine a situation in which you would approve of a policeman striking an adult male citizen” (Global support, p. 656).  Support for “reasonable force” was assessed with questions such as one that asked whether they supported striking someone attempting to escape custody.  Support for “excessive force” was assessed with questions such as their approval of striking a citizen who used vulgar or obscene language with a police officer. 

Consistent conservatives were more supportive of police use of force (global, reasonable, and excessive) than conflicted conservatives. Indeed, conflicted conservatives, on some measures, were indistinguishable from moderates or liberals.   This is not an issue of simply how conservative they are.   Self-described ‘extremely conservative’ and ‘conservative’ respondents did not differ in their global views of police use of force or in their views of police use of reasonable force from those who described themselves as slightly conservative.  Those who described themselves as slightly conservative were, however, less supportive of police use of excessive force than the extreme conservatives. For conflicted conservatives, but not consistent conservatives, global support for police use of force was associated with confidence in government.

Conclusion: It appears that conflicted conservatives are a distinct political group, looking much more like liberals and moderates than consistent conservatives in their level of support for police using force.  Said differently, understanding conservatives’ criminal justice attitudes requires more information than simply where on the political spectrum a respondent is.  Those conservatives who favour government spending on important social issues such as health are very different in their views of the police use of force from those who oppose this form of funding. 

Reference: Silver, Jasmine R. and Justin T. Pickett (2015) Toward a Better Understanding of Politicized Policing Attitudes: Conflicted Conservatism and Support for Police Use of Force.  Criminology, 53(4), 650-676. 

Back to top

Publication Type

Volume Number


Issue Number