Criminological Highlights Vol. 6, No. 4 – June 2004

View this issue as a PDF: PDF iconCrimHighlightsV6N4.pdf

This issue of Criminological Highlights addresses the following questions:

  1. Can community drug treatment programs save taxpayers’ money?
  2. Why do black Americans commit homicide at higher rates than whites?
  3. Does everyone “age out” of crime?
  4. Can the practice of “racial profiling” be justified as being “efficient policing”?
  5. Why does the employment of young adults only sometimes reduce crime?
  6. Why does marriage appear to contribute to a reduction in crime?
  7. Are judges likely to impose disproportionately harsh sentences when given the opportunity to do so?
  8. How can we account for the anomalously high rate of imprisonment in the U.S.?

While community drug treatment programs for people released from prison are not a panacea for crime, they may be effective enough to be justified solely on the basis of reduced incarceration costs in the future.   

Background.  It is generally recognized that many prison inmates also have substance abuse problems.  Aside from other negative consequences that may result from addiction, it has been shown (e.g., Criminological Highlights, 6(3) #6) that drug use leads to higher rates of property crime. Unfortunately, drug treatment in prison alone is not likely to be as effective as programs that continue in the community after release. In fact, some research has suggested that well-run therapeutic community programs which focus on reducing the use of illegal drugs as well as antisocial behaviour - while simultaneously developing employment skills and pro-social attitudes and values - appear to lower not only drug use but also recidivism. 

This study examines the impact of an intensive (6 month) therapeutic community work release program in Delaware. Prisoners re-entering society were randomly assigned to either the standard work release program that existed in the Delaware correctional system or the more intensive therapeutic community work release program for drug-addicted offenders. Several individuals in each group had also received an in-prison addiction program. Additionally, a number of the standard work release inmates received some counseling in other regular, more broadly available programs. 

The results focused on a “cost effectiveness” measure which looked only at the (relative) treatment costs in comparison to the direct expenditures incurred or avoided by corrections. In an 18-month (547 day) follow-up period, the treatment group spent significantly fewer days in prison than did the comparison group (74 vs. 104 days). Further, several participants within the treatment group also received a 6-month aftercare program (involving group and individual counseling sessions as well as urine testing for drugs). These individuals spent, on average, even fewer days in custody than those in the treatment group who did not receive this aftercare intervention (43 vs. 92 days).  

The estimated cost of the therapeutic work release program was US$1937 per offender. As this intervention apparently saved approximately 30 days in prison, the average treatment cost would be roughly US$65 per day compared to an estimated imprisonment cost of US$57. Taken alone, these incarceration expenditures might lead one to conclude that the program was not cost effective. However, a number of factors were not included in the analysis - prosecution costs of those who might have re-offended during the 18-month period, criminal justice costs after 18 months, and additional social benefits (e.g., less drug use, more stable employment, less dependence on the social welfare system). The calculations also assume that there were no treatment costs for the comparison group since it was not possible to estimate them. Moreover, the savings for those who participated in the program as well as the aftercare component were even more favourable: for every US$19 invested in the program plus aftercare, 1 day of incarceration was avoided. 

Conclusion. This analysis suggests that a carefully implemented, comprehensive and intensive treatment program for drug addicted ex-prisoners can pay for itself solely in terms of reduced subsequent incarceration costs. However, it does not mean that all treatment interventions are “cost effective” (see Criminological Highlights, 1(6) #2). Nevertheless, this study serves as a reminder that despite their limitations, treatment programs may be good investments in reducing future criminal justice costs.

Reference: McCollister, Kathryn E.; M.T. French; J.A. Inciardi; C.A. Butzin; S.S. Martin and R. M. Hooper (2003). Post-Release Substance Abuse Treatment for Criminal Offenders: A Cost-Effectiveness Analysis. Journal of Quantitative Criminology, 19, 389-407.

Back to top

The difference in the homicide rate for blacks and whites in U.S. cities is better explained by the differences in resources (i.e. income, education, professional jobs) available to the two groups than by their relative level of serious disadvantage (i.e. poverty rate, male joblessness rate, proportion of female-headed families). 

Background. In most U.S. cities, the rate at which black Americans commit homicide exceeds that of whites, with a mean difference between blacks and whites of 37.7 homicides per hundred thousand residents. However, the range across different cities is enormous. More specifically, the difference between black and white rates of lethal violence ranges - in 126 central cities - from –8 to +104 (per hundred thousand residents). These divergences have often been explained in terms of differences in the rates of disadvantage of the two groups – e.g., differences in the proportion of black minus the proportion of white residents below a poverty line or excluded from the labour force. An alternative formulation would be to examine the resources available to each group. For example, one might look at differences between the two populations in terms of household incomes or the proportion of college graduates and individuals employed in professional or managerial positions.  

This study examines the relative importance of these two indices – disadvantage and resources - for the size of the difference in the rates of homicides for blacks and whites.  Interestingly, the correlation between black-white differences in rates of disadvantage (i.e. poverty rates, female-headed families, and male joblessness) and black-white differences in rates of resources (i.e. median income, college graduates, professional jobs) is moderate, but not overwhelmingly high. For instance, the correlation between the size of the gap in median income and the difference between the proportion of families living below the poverty line is a moderate 0.43, suggesting that although the two measures are correlated, they represent distinct constructs. 

The results suggest that when certain other characteristics of the cities (e.g., level of segregation and differences in home ownership rates - a proxy measure of community stability) are held constant, the differences in the rates of “disadvantage” do not predict the gap in the homicide rates for blacks and whites. In contrast, the three measures of the difference in overall resources available to the two groups (i.e. median income, proportion of college graduates, proportion of professionals) significantly contribute to an understanding of race differences in homicide offending. In addition, the measures of racial segregation and relative rates of home ownership existing in the community were consistent predictors of the relative rate of homicide. 

Conclusion. The findings of this study support the conclusion that “race differences in [homicide] are affected by substantial gaps between blacks and whites in community stability” (p.660). Further, “it becomes apparent when the actual racial homicide gap is studied that levels of black and white killing diverge so much because these groups are positioned in such dramatically different ways with respect to the core factors reflecting traditional crime” (p.665). “When racial residential segregation is higher, and whites’ level of home ownership, median income, college graduation, and professional workers exceed those for blacks to a greater degree, the racial gap in lethal violence is wider” (p.664).

Reference: Vélez, María B.; Lauren J. Krivo and Ruth D. Peterson (2003). Structural Inequality and Homicide: An Assessment of the Black-White Gap in Killings. Criminology, 41, 645-671.

Back to top

In a longitudinal study of delinquent males followed from age 7 to age 70, there were no identifiable groups whose rates of offending did not decline with age.

Background. The belief that a group of offenders exists whose crime rate does not drop with age is attractive to policy makers. It suggests that criminal activity can be substantially reduced if something were done (e.g., treatment or incapacitation) with such a group that would stop future offending. Within this context, policy makers and “criminal careers” researchers have focused on identifying “the subset of chronic offenders known as serious, violent offenders” (p.558) (See also, Criminological Highlights, 1 (3) #7).

This study followed a sample of 500 boys born between 1924-1931 through to 1996. Originally part of a study carried out by Sheldon and Eleanor Glueck, these boys were subsequently traced until the 1990s by way of various death and (state and federal) offending records. The main focus of this research was to determine whether the age-crime relationship – an increase in crime from childhood into adolescence or early adulthood and a decline thereafter – was consistent across groups of offenders.

The results are easy to describe.

  • Pooling across all offences, there was a sharp increase in offending which peaked in adolescence and was followed by a slower decline throughout the life course.
  • For property offences, the rate peaked in adolescence, with a very sharp rate of decline immediately afterward. In fact, the rate of (property) offending was quite low by age 30.
  • For violent offending, the peak was in the 20s, and the drop-off in offending rates was substantially slower, “with some offenders remaining active well into their 40s…” (pp. 565-6).
  • Drug and alcohol offending was relatively constant between age 20 and approximately age 47 before it sharply declined. 
  • In an attempt to identify groups that might not show a decline in offending, 13 measures from the boys themselves (as youths), their parents, official records, and teachers were used. More specifically, this study examined differences in such variables as IQ, age of onset of misbehaviour, psychological assessment indicators of the boys and level of delinquency in adolescence. Youths were grouped into those 20% with the most “risks” vs. the rest. The same pattern emerged – early peaks in late adolescence for property crimes, and later peaks (late 20s) for violent offences, and a flatter curve peaking in the 30s for drug/alcohol crime. Though obviously the rate was higher for the “high risk” offenders, the general shape of the curves was the same. 
  • High rate chronic offenders showed the same pattern of drop-off in offending demonstrated by those with lower rates of offending. 
  • Even when the high rate offenders with high “family risk” factors were compared with the remaining youths, the results were identical: the drop-off in offending occurred more or less at the same time.

Conclusion. “The data are firm in signaling that persistent and frequent offending in the adult years is not easily divined from zeroing in on juvenile offenders at risk” (p.577). There is variation in the age at which certain groups of men will peak in their offending rates, but there are no identifiable groups whose level of offending does not drop off with age. “Crime declines with age even for active offenders” (p.585). As such, it was impossible to find a “life-course persistent group [that] can be prospectively or even retrospectively identified based on theoretical risk factors at the individual level in childhood or adolescence” (p.588). 

Reference: Sampson, Robert J. and John H. Laub (2003). Life-Course Desisters? Trajectories of Crime Among Delinquent Boys Followed to Age 70. Criminology, 41, 555-592.

Back to top

When stopped by the police, blacks and Hispanics in the U.S. are more likely than are whites to receive a traffic ticket, be arrested, or be subject to the use of force.  However, they are not more likely to be carrying contraband.

Background. “The practice of targeting racial minorities for routine traffic and pedestrian stops [in the U.S.] can be traced back to the war on drugs, which promoted profiling as an effective policing tactic to detect drug offenders” (p.50). At the height of this American era, “drug arrest rates were five times higher for blacks than for whites despite consistent evidence… of similar rates of drug usage by the two racial groups…” (p.52). Indeed, these types of findings have challenged the notion defended by profiling advocates that “it is reasonable and efficient for police to consider race in their decision making [on whom to stop]” (p.54). However, in order to conclude that the over-representation of black citizens among those stopped does not, in fact, constitute efficient policing, it is necessary to know the “hit rate” on these stops – the proportion of searches which produce contraband. The research on this issue is mixed, with 8 of 16 published studies showing higher “hit rates” for black and/or Hispanic citizens than for whites. Unfortunately, these studies have typically been carried out in single cities or states. As such, it is not known the extent to which the findings may be generalized.

This study used data from a national survey of U.S. residents. Overall, black drivers were somewhat more likely than whites to have been stopped in the previous year. Further, blacks who were stopped also had a higher probability than white drivers of being given a ticket, searched, arrested, and subject to the use of force by the police (including the use of handcuffs). This relationship held even after the researchers statistically controlled for various characteristics of the driver (e.g., age, income), the stop (e.g., its stated purpose - an impaired driving assessment, traffic offence, vehicle defect) and the community (e.g., the location in which the stop took place - the centre of a large city, a non-central area). 

More importantly, although blacks and Hispanics who were stopped were more likely to be subject to various police actions, “contraband was discovered on fewer minority drivers than on white drivers…. [Specifically,] 16% of the Caucasian drivers who were searched were found to be in possession of contraband, compared to only 7.5% of non-Caucasian drivers…” (p.76). Clearly, these findings are “[c]ontrary to the argument made by many law enforcement officials that minorities are more likely to be carrying drugs and/or weapons” (p.76). 

Conclusion. The two major findings of this study – that blacks and Hispanics who are stopped are more likely to be subject to police actions but less likely to be found to be involved in any criminal wrongdoing – suggest that “targeting drivers solely or even partially on the basis of their race/ethnicity is not an effective, efficient, or responsible policing strategy at the national level [in the U.S.]” (p.82). It would certainly seem that it is not the drivers but the practice of racial profiling that needs to be stopped. 

Reference: Engel, Robin Shepard and Jennifer M. Calnon (2004). Examining the Influence of Drivers’ Characteristics during Traffic Stops with Police: Results from a National Survey. Justice Quarterly, 21, 49-90.

Back to top

Employment for young adults reduces crime largely because previously delinquent youths not only come into contact with prosocial peers in the context of their job but also stop associating with those who supported their delinquent activities.

Background. Various life events – such as employment or marriage – appear to be related to reductions in criminal behaviour among young adults. There are at least two possible explanations for the effect of these changes. First, they may be associated with “an extensive set of obligations, expectations, and interdependent networks that can facilitate social control” (p.184). Alternatively, “prosocial work relations may restructure friendship networks by diminishing contact with delinquent peers” (p.185). More specifically, “entrance into a job introduces an employee to a new set of peers… [and the] quality and content of the relationships that occur between coworkers, not necessarily with an employer or the institution of work, may determine [the impact on crime]” (p.187). This latter explanation would be consistent with findings demonstrating that employment is often associated with an increased risk of offending for adolescents as youth coworkers are, themselves, not frequently committed to conventional values. 

This study used data from the U.S. National Youth Survey. In addition to various measures of the youth’s commitment to work, the survey included a measure of the respondent’s assessment of the degree of disapproval that would be felt by his/her coworkers upon discovering that he/she was engaging in any criminal behaviour. The main dependent variable was a self-report measure of offending. Youths were interviewed when they were 15-21 years old and again three years later when they were young adults (ages 18-24). When the researchers controlled for the level of offending at the first interview, they found that having “prosocial co-workers” and having a long period of stable employment were associated with lower levels of involvement in crime. Similar results were found with respect to drug use. More specifically, reduced rates of drug use were shown to be related to having prosocial coworkers, and especially, having stable employment with prosocial coworkers.  

Conclusion. Contrary to the view “that employment builds informal social control by establishing [successful association] with employers” (p.198), these results suggest that “employment may set in motion a process that reduces offending by affecting peer networks” (p.198). For these young adults, stable employment - on its own - was associated with reduced levels of criminal activity. However, the presence of prosocial coworkers was also directly related to lower rates of offending and drug use. Therefore, it would appear that employment alone may not be a central factor in the reduction of crime as much as is employment in settings in which coworkers are not likely to be involved in, or approve of, crime.

Reference: Wright, John Paul, and Francis T. Cullen (2004). Employment, Peers, and Life-Course Transitions.  Justice Quarterly, 21, 183-201.

Back to top

Peer networks, marriage and other significant life events only partially explain a reduction in crime among those who were serious offenders in their youth. The impact of these factors also depends on the ways in which these changes fit into an individual’s views of the kind of person that he/she wants to be.  

Background. It is well established that most young offenders cease criminal activity in early adulthood. Research on this “desistance” process has often suggested that marriage as well as other significant life events are important factors in explaining this reduction in offending. Similarly, the act of turning away from delinquent friends has also been found to be associated with diminished criminal behaviour. However, it may equally be that marriage and peer relations are linked whereby marriage may be important in understanding changes in friendship relationships which, in turn, affect offending. The difficulty with such structural analyses is that they do not take into account the manner in which people alter their views of themselves, particularly in terms of their identity as a ‘non-offender’. Indeed, one’s choices of partners and friends may reflect – rather than cause – changing perspectives of oneself.

This paper reports the results of a longitudinal study of men and women initially interviewed when they were incarcerated as young delinquents in Ohio. Thirteen years later, 85% of them were located (both in and out of prison) and were asked questions about the nature of their lives in the intervening period. Detailed information was collected – and verified against official sources when possible - on respondents’ levels of criminal involvement as well as those of their friends and spouses. Additional questions were also asked about their peer networks –those people with whom they spent time and the degree of influence that these individuals had on them.

Within this sample of (originally) serious delinquents, marriage per se was not significantly related to the level of criminal involvement of their friends. Rather, the more relevant factor seemed to be whether their spouses supported or discouraged criminal activity. Indeed, the level of offending of the respondent’s spouse was a relatively strong predictor of his/her own level of criminal involvement. Compared to their responses when they had been adolescents, people also indicated that peers influenced them less as adults. Further, people in adulthood tended to make conscious decisions about the type of friends that they wanted to have. Some participants claimed that they actively avoided people whom they knew were likely to commit offences while others maintained relationships with delinquent friends but were able to avoid engaging in crime because of perceptions about their own identity as non-offenders. Clearly, the life choices that people were making about friendships – more than attachment to one’s spouse or job stability – appeared to be important in predicting adult offending. 

Not surprisingly, women showed lower levels of involvement in crime than did men at the time of the follow-up interview. However, this difference can be explained by the fact that male respondents’ friends tended to have higher levels of criminal activity. In addition, women seemed to be less influenced than men by the illicit behaviour of their spouse or friends. 

Conclusions: This paper suggests that it may be worthwhile to rethink our assumption that certain life course events - such as marriage - are likely to have automatic effects on the probability that a person will cease to offend. The manner in which these events are interpreted appears to be crucial in understanding their effect on crime. Similarly, the impact of associations with previously delinquent peers can and does change over time - people become less influenced by the misbehaviour of those with whom they associate. As such, conditions imposed on offenders that prohibit contact with prior offenders should be considered in light of the somewhat complex manner in which these friendships are related to criminal behaviour.

Reference: Giordano, Peggy C.; Stephen A. Cernkovich and Donna D. Holland (2003). Changes in Friendship Relations over the Life Course: Implications for Desistance from Crime. Criminology, 41, 293-327.

Back to top

When given the opportunity to impose sentences that were explicitly longer than would be considered proportionate to the gravity of the offence, judges in Victoria, Australia largely declined to do so, despite the popularity of these provisions with the public and politicians.

Background.  Previous research in England on the use of dangerous offender legislation by judges has suggested that “there was a judicial reluctance to apply the full force of provisions such as longer than commensurate sentences and automatic life sentences” (p.82). Generally speaking, “proportionality has been regarded as a fundamental tenet of the common law by the High Court of Australia” (p.82).  Nevertheless, the Victoria (state) government introduced legislation in 1993 that required the court “to regard protection of the community as the principal purpose for which the sentence is imposed”. In addition, it was stipulated that “in order to achieve this purpose, the court may impose a custodial sentence ‘longer than that which is proportionate to the gravity of the offence considered in light of its objective circumstances’”(p.83). It was further permitted that anyone over age 21 and found guilty of one or more of 50 “serious” offences could be sentenced to an indefinite term of imprisonment, notwithstanding the fact that other maximum sentence lengths would normally apply. Clearly, this legislation was a threat to the principle of proportionality in sentencing. 

Systematic examination of the effect of legislation such as this one has detected “a juridical tradition… which is generally resistant to such policies” (p.85). For instance, judges will limit the application of these laws to exceptional cases. Similarly, they will interpret the legislation in such a way as to restrict its use. In this study, researchers examined 553 cases appearing before the court between 1994 and 2002 in which the application of these special provisions was considered.  These cases largely involved serious sexual offences whose rate of guilty pleas was lower than that for the court as a whole.  

The experience in the courts in Victoria suggests that indefinite sentences are not imposed even when they could be. In fact, only 4 of these 553 cases resulted in such a sanction, and in only 11 additional cases were sentences handed down that were longer than proportionate, as allowed by this law. Only 5 of these sentences were upheld after appeal and all were for serious sexual offenses. Further, in those cases in which a disproportionate sentence was not imposed, judges typically indicated that a proportionate sentence would provide sufficient protection to the public. Interestingly, this judicial resistance to indefinite sentences appears to have also influenced prosecutors who have made relatively few applications for this type of sanction.  

Conclusions. “Judicial ownership of sentencing and the fierce way in which judicial discretion is defended by the judiciary may go part of the way to explain the reasons for why judges employ avoidance tactics so as not to apply provisions they do not agree with. But such reluctance may also be due to their recognition that the sentences that they are urged to impose will not protect the community, regardless of prevailing popular opinion” (p.98). 

Reference:  Richardson, Elizabeth and Arie Freiberg (2004). Protecting Dangerous Offenders from the Community: The Application of Protective Sentencing Laws in Victoria. Criminal Justice, 4, 81-102. 

Back to top

Why does the U.S. imprison so many people?  Cycles of crime, sensibilities about criminal activity, moral panics, and governmental structures that allowed unthinking and unthinkable “tough on crime” policies to flourish coincided to create strategies whose purpose was to increase levels of imprisonment.  

Background. American imprisonment rates are exceptional. In the 1930s, they were comparable or lower than those of many European countries. Even until the 1970s (when the U.S. rate was approximately 144 per 100,000 residents), the difference between the U.S. and elsewhere was not substantial. In contrast, the current level of incarceration in America (700 per 100,000 residents) is dramatically higher than that in any western country. In explaining this phenomenon, it is easy to see what does not account for it:

  1. Crime rates.  The U.S. crime rate is exceptional only in terms of gun violence. Further, while levels of criminal activity have also increased in other countries (e.g., Germany and Finland), imprisonment rates have not. 
  2. Public opinion.  In their views of crime and punishment, American citizens do not appear to be unique. Like others (e.g., Canadians), they will accept - if given the choice - less harsh policies to deal with offenders.  
  3. Partisan politics. Crime is a relatively new political issue and was used by Republicans to lure traditional Democrats to the political right. But why only in the U.S.? Crime has been part of the politics of other countries in recent decades (e.g., in Canada) but has not had the same impact. 
  4. Political Reconfiguration.  The need for the state to be seen as capable of solving important problems in an era in which the public has lost faith in government and the electorate is split on the basis of single issues has also been offered as a possible explanation.  However, we are left with the same question as to why only the U.S. has responded to these common problems of western democracies with policies to increase imprisonment.
  5. Thinking about “risk”. Though there may have been a governmental shift toward policies based on the assessment and reduction of “risk,” such thinking is not unique to the U.S. (or the U.K., in which imprisonment is also increasing). As such, it is not useful as an explanation. 
  6. Postmodern angst brought about by an increase in crime, decreased confidence that liberal policies can contain crime, and perceived increased vulnerability to crime by the traditionally liberal elite has also been proposed to explain U.S. incarceration rates. However, for this account to be true, one must again argue – unconvincingly – that such experiences only occurred in the U.S. (and, perhaps, the U.K.).

Several alternative explanations are offered for the high rates of imprisonment in the U.S. today. First, the state bureaucratic apparatus - including the judiciary - is relatively powerful and autonomous in much of Europe and less subject to political pressures. Second, it is suggested that general crime policies – like those specifically related to drugs – have emerged as a response to a cultural climate in which any approach which was not (perceived as) punitive in nature was rejected because it was seen as ‘soft on crime’. For instance, drug use became associated with the poor and minorities in the public’s mind during the period in which drug issues were salient in the U.S.. Hence, drugs were consequently seen as being dangerous and/or evil, rendering punitive drug policies easy to exploit politically. As a result, although (drug) crime rates had already begun to reverse, repressive strategies were still implemented as the most ‘viable’ political response. Finally, the recent increase in the U.S. crime rate (1960-1990) occurred within a context of a much longer overall decline. This increase coincided with a series of short-term moral panics in the U.S. during which “problems are exaggerated, public attitudes are polarized, traditional values of moderation and balance are cast aside, moral entrepreneurs become influential, and extreme policies are adopted” (p.98). Not surprisingly, cultural sensitivities changed, with people becoming less sympathetic to the offender and more moralistic and judgmental about wrongful behaviour.

Conclusion. In the past 30 years, “cycles, sensibilities, and moral panics coincided in ways that produced current crime control and penal policies. Crime rates rose… the world changed… [and] these things raised fears and anxieties that were in part displaced to people (criminals) and things (crime and disorder) that were ready objects of hatred and derision” (p.138). 

Reference: Tonry, Michael (2004). Thinking about Crime: Sense and Sensibility in American Penal Culture. New York: Oxford University Press. Chapters 2 and 5: “Why So Many Americans are in Prison” and “Crime Trends and the Effects of Crime Control Policies.”  

Back to top

Publication Type

Volume Number


Issue Number