Criminological Highlights Vol. 12, No. 3 – January 2012

View this issue as a PDF: PDF iconCrimHighlightsV12N3.pdf

This issue of Criminological Highlights addresses the following questions:

  1. Are long sentences more effective than shorter sentences in reducing reoffending?
  2. Why do western countries vary in their imprisonment rates?
  3. Do intensive foot patrols reduce crime?
  4. Does being apprehended and arrested for a crime act as a deterrent?
  5. Why aren’t people deterred by harsh sentences?
  6. Will new and better community sanctions reduce imprisonment?
  7. How is a period of imprisonment more punishing for Blacks than Whites?
  8. Do trustworthy looking people have an advantage at trial?

The length of time an offender spends in prison on the first prison sentence has no discernible impact on the likelihood that he or she will reoffend.

There are theoretical reasons to believe that the time that an offender spends in prison could either increase or decrease the likelihood of reoffending. If time in prison were to convince offenders that the risks of offending are too high, long prison sentences could reduce offending.  Alternatively, long periods of time in prison could increase subsequent offending by reinforcing deviant values, stigmatizing the offender, and/or making it more difficult for the offender to obtain legitimate employment upon release.  

This study investigates the impact of the length of time in prison on reoffending in a three year period after release for a group of Dutch offenders sentenced to prison for the first time in 1997.  All were under 40 years old, and were convicted of violent, property, or drug offences. In order to control for differences between those getting longer and shorter sentences, two somewhat independent techniques were used.  First, the 4,683 offenders were divided into four groups according to the best estimate of their predicted ‘trajectories’ of offending at the time of sentencing. Second, pairs of offenders receiving ‘long’ and ‘short’ sentences were matched on various other measures (age, sex, whether the offender was an immigrant, 9 different measures of past criminal convictions, and various measures of the seriousness of the offence for which they were being sentenced).  In the end, 4,096 offenders were successfully matched.  Those excluded were largely those with extremely long or short sentences.  Most importantly, the matched pairs were always of the same age and sex and were in the same offending trajectory group.

Offenders were divided into 5 groups according to the time that they served, running from ‘less than one month’ to ‘more than one year’.  Dutch prison terms are short as compared to the US, but comparable to those in Northern Europe and Canada. 86% of the sentences in this sample were under a year, a figure which is comparable to overall Canadian sentences (89% under 1 year). 

The findings are easy to describe. When adequate controls were imposed on the comparisons, pairs of similar offenders with different sentence lengths did not differ in reoffending.  Two measures were used: the felony reconviction rate and the proportion reconvicted (one or more times) within three years.  Essentially, the data show that the length of time in prison (ranging from under a month to over a year) had no effect on reconviction.  It is important to note, however, that without any controls, those receiving long sentences looked somewhat less likely to reoffend.  It is easy to understand why: those receiving long sentences were very different from those receiving shorter sentences on many dimensions related to reoffending. What is important, however, is that when age, offending trajectory and a large number of other important controls are introduced, there was essentially no consistent impact of time in prison on offending.  Said differently, when cases that are similar on relevant dimensions are compared, time in prison has no discernible impact on reoffending. 

Conclusion: Previous research suggests that sending an offender to prison rather than imposing a community punishment may be criminogenic (see Criminological Highlights, 11(1)#2).  For those who are imprisoned for the first time, the length of time in prison appears to be irrelevant to future offending.  Obviously prison sentence length is going to vary for reasons other than likelihood of reoffending (e.g., for the purpose of achieving proportionality).  These data, however, suggest that judges, when sentencing an offender to a first prison sentence, should not vary the sentence length because of a belief that sentence length affects reoffending.  

Reference:  Snodgrass, G. Matthew, Arjan A. J. Blokland, Amelia Haviland, Paul Nieuwbeerta, and Daniel S. Nagin (2011).  Does the Time Cause the Crime?  An Examination of the Relationship Between Time Served and Reoffending in the Netherlands. Criminology, 49(4), 1149-1194.

Back to top

National imprisonment rates vary dramatically but, generally speaking, they reflect the political and social values of each country. 

Imprisonment rates of European and English-speaking countries vary dramatically.   According to 2007 imprisonment data, the rates in Nordic countries are very low and vary from 55 prisoners per one hundred thousand residents in Iceland to 78 in Sweden.  At the other end of the spectrum, the Eastern European countries have relatively high rates (e.g., 212 in Poland, 206 in the Czech Republic, 319 in Latvia), but not as high as that of Russia (592) or the USA (760).  This study looks at imprisonment rates in 30 countries in Europe and three countries (Canada, Australia, New Zealand) outside of Europe.  The US and Russia were not included in the analysis. 

Generally speaking, reported crime, victimization rates, and the rate of non-traffic criminal convictions were unrelated to imprisonment rates.  Homicide rates were positively correlated with imprisonment rates, but to a large extent this was due to three outliers (Latvia, Estonia, and Lithuania), all of which had high homicide rates and high imprisonment rates. If the other 27 countries are looked at, there is very little relationship. But to suggest that this may be a causal relationship between high homicide rates and high imprisonment rates ignores the possibility that a third factor – such as social policies - may be responsible for both high homicide rates and high imprisonment rates. 

When looking at social policy indicators, the picture is quite consistent.  Those countries with relatively low levels of economic disparity have relatively low imprisonment rates.  Those countries that spend more funds on social welfare concerns (expressed as percent of gross domestic product or in the number of Euros spent per person on social welfare matters) tend to have low imprisonment rates.  Similarly, those countries that score high on measures of actual welfare, well-being, and prosperity in various areas of life have low imprisonment rates. Those countries in which citizens indicate that they have high levels of trust in their fellow citizens, the police, the justice system, and their parliament tend to have low imprisonment rates. Finally, countries in which the national government operates largely by consensus – in contrast with a simple ‘winner takes all’ majority rule – tend to have low imprisonment rates.

As with research on individual preferences for high imprisonment policies, countries in which large portions of the population feel unsafe and with high proportions of people who report intolerance toward  various groups, including offenders, tend to be countries with high rates of imprisonment. 

Conclusion: Imprisonment rates should not be seen as being an accident of a country’s criminal justice policies; they appear to reflect values and approaches to governing within each country.  “Prison rates (and social policy) are an outcome of policy choices and political actions” (p. 318).  “Moderate penal policies have their roots in a consensual and corporatist political culture, in high levels of social trust and political legitimacy, and in a strong welfare state…. Consensual politics lessens controversies, produces less crisis talk, inhibits dramatic about-turns and sustains long-term consistent policies. In other words, consensual democracies are less susceptible to political populism” (p. 321-2). 

Reference: Lappi-Seppälä, Tapio (2011).  Explaining Imprisonment in Europe.  European Journal of Criminology 8 (4), 303-328. 

Back to top

Very intensive foot patrols by police can have an impact on street crime. 

Police foot patrols have been seen as a popular way to address crime, though the evidence that they actually deter crime has been weak. The public appears to believe that if there is an officer on foot patrol in their neighbourhood, they will be safe.  From a management perspective, foot patrols are expensive. If, however, foot patrols are used selectively to target crime “hotspots” – locations where crime rates (or street crime in particular) are high – it has been suggested that they may be especially effective.  

Unlike some studies that looked at relatively large geographic areas, this study examined the impact of intensive police foot patrols on street crime using a large number of small geographic areas.  Crime hotspots were identified, in early 2009 in Philadelphia, by looking at the number of homicides, aggravated assaults, and robberies that had occurred outdoors in recent years.  In all, 120 hotspots were located, each including at least one of the most violent street corners in the city. These hotspots had an average of 14.7 intersections and 1.3 miles of streets. These 120 hotspots were then divided into 60 pairs of hotspots with similar numbers of violent incidents.  One of each pair was then randomly determined to be a ‘control’ hotspot (with no special change in police patrol intensity). The other received intensive patrols for 12 weeks in addition to normal policing.  The intensive patrols consisted of 2-person foot patrols for 12 weeks from 10a.m. until 2 a.m.,  5 days a week (Tuesday morning to early Sunday morning).  In all, then, 57,600 hours of 2-person police patrol (115,200 person-hours) were used during the 12 week period in the 60 intensive patrol hotspot areas. The activities of the police officers varied considerably across areas in terms of the number of recorded pedestrian and vehicle stops, arrests, and recorded disturbances and drug-related disorder. 

Overall, there was a slight reduction in the average number of violent crimes recorded in the experimental areas, compared to the average number before the intensive foot patrols (a reduction of about 0.88 crimes per area during the 12 week period).  In the control areas, there was a slight increase in the number of crimes during the ‘treatment’ period, as compared to the earlier period (0.52).   However, this apparent relative reduction only occurred in the highest crime areas. These were the areas, not surprisingly, in which the foot patrol officers were most likely to have direct contact with citizens as a result of arrests or responding to various forms of disorder.  In the relatively low crime areas (which were, of course, hotspots relative to the city as a whole), the patrols had essentially no impact on crime.  

However, it would appear that some of the violent crime reduction in the intensive foot patrol areas was a result of displacement to adjacent areas.  It was estimated that 90 violent crimes were averted in the target areas as a result of the intensive foot patrols, but an estimated 37 of these crimes were displaced to adjacent areas. Thus there were an estimated 53 fewer crimes as a result of the intervention, or one crime for every 1087 hours of 2-person patrols (or 2174 person-hours of patrol). 

Conclusion:  It would appear that highly intensive policing can modestly reduce the number of violent crimes that take place in an area.  In part because the effect is small and is limited to the very highest crime areas, it is difficult to know whether to attribute the drop in crime to the mere presence of a police officer in the area or to the activities of the police in the neighbourhood. The data would suggest that it may be that foot patrols can only deter violent street crime in very violent areas.  However, the investment of police time for each crime averted was non-trivial.  

Reference: Ratcliffe, Jerry H., Travis Taniguchi, Elizabeth R. Groff, and Jennifer D. Wood (2011). The Philadelphia Foot Patrol Experiment: A Randomized Controlled Trial of Police Patrol Effectiveness in Violent Crime Hotspots. Criminology, 49(3), 795-831.

Back to top

When youths are apprehended and arrested for offences, their perceptions of the likelihood of being caught in the future increase – but not very much. 

Much of the popular and academic interest in deterrence has to do with general deterrence, or reductions in crime (by people other than the offender) through harsher penalties. General deterrence, however, has been shown largely to be ineffective. But punishments may be effective in other ways.  Specifically, it may be that catching and arresting people for offences will reduce their future offending by increasing their assessments of the likelihood that they will be caught and arrested should they offend in the future.  In other words, a criminal justice system that is good at catching offenders may teach them, in effect, that crime does not pay.  This study looks into this possibility with a sample of adolescents who had been found guilty of relatively serious offences in either of two U.S. counties.

These adolescents were interviewed once every 6 months for three years starting when they were, on average, about 16.5 years old.  Among other things, they were asked how likely it was that they would be caught and arrested if they were to commit each of seven different crimes ranging in seriousness from ‘stealing clothes from a store’ and ‘vandalism’ to ‘robbery with a gun’ and ‘stabbing someone’ (p. 652). They were also asked to report how many times, if any, they had committed each of 22 offences.  Arrests were recorded from juvenile court records in the two locations.  The focus of the study was on the youths’ estimates of the probability of being apprehended as a function of whether they had been caught for any offences they had committed during this period. 

Overall, the findings showed that the youth’s estimate, during any six month period, of being apprehended for offending was a function of two things: the youth’s perception of being apprehended prior to that period and whether the youth had been apprehended for offending during the previous six months.  Overall, if a youth committed a crime, the youth’s estimate of being apprehended increased by 6.3% if the youth had been arrested compared to if they had not.  It would appear that arrests for one type of crime (aggressive crimes) also affected respondents’ perceptions that they would be apprehended for income-generating offences, though this effect is slightly smaller. In other words, there was some evidence that the impact of an arrest was not crime specific.  Overall the data show that although the youths did change their subjective estimate of being apprehended, there was a good deal of variability in whether and how much updating of these estimates actually took place. 

Conclusion: It appears that “even among serious offending juveniles, an arrest still has a potential deterrent effect, at least as far as increasing risk perceptions. However, among more experienced or frequent offenders, this gain from deterrence may be reduced or, in some cases, lost all together” (p. 691).  There was, however, a great deal of individual variability. Thus it cannot be assumed that apprehension and arrest is, for all youths, a crime reducing strategy.  It is difficult, moreover, to estimate how much impact the changes in perception (of apprehension) may have on actual offending. One study found that a 10% change in the perceived likelihood of apprehension reduced offending by approximately 3% to 8% depending on the offence. Applying these findings to the present results would suggest that the impact of an arrest would be quite modest – reducing offending through individual deterrence by between 1.2% and 3.2%.     

Reference: Anwar, Shamena and Thomas A. Loughran (2011). Testing a Bayesian Learning Theory of Deterrence Among Serious Juvenile Offenders. Criminology, 49 (3), 667-698.

Back to top

Most active and violent offenders don’t think that they will be caught or have no idea what punishment to expect from their crimes if they were to be caught.  More severe sentences would, therefore, have no impact on their likelihood of offending. 

Those who suggest that harsher sentences would reduce crime appear to endorse the economic model that, in general, potential offenders make decisions that are informed by evidence of the consequences of what would happen to them if they were apprehended. However, for the consequences (e.g., the severity of the sentence) to make a difference, the offender, at the time of the offence, must both be thinking that there is a plausible chance of being caught and  know what the likely punishment would be. 

In this study, 278 prison inmates were asked questions about the offence that got them in prison.  Specifically, they were asked “When you committed this crime, how likely did you think it was that you would be caught?” This was answered on a 4 point scale ranging from ‘very likely’ to ‘I did not think I would get caught’ plus the alternative ‘I did not think about it.’   They were also asked “When you committed the crime, did you know what the likely punishment would be if you were caught?”  Again, a four point scale was used ranging from ‘I knew exactly what the punishment would be’ to ‘I had no idea or thought I knew but was wrong’, plus the alternative ‘I didn’t think about it’ (p. 303).  By interviewing only those apprehended and punished rather severely, one obviously misses those who were not apprehended or imprisoned.   However, the purpose of this study is “to determine to what extent current offenders could be dissuaded by more severe sentencing” (p. 301). 

Overall, 42% of the 278 offenders indicated that they did not think about whether or not they would be caught and an additional 34% did not think they would be caught or thought it was not likely.   When asked what they thought the punishment would be at the time they were committing the crime, 35% indicated that they didn’t think about it, and an additional 18% had no idea or later found out that they were wrong.  In total 76% of these prisoners were oblivious to the fact that they might get caught or what the penalty would be (or both).  Raising penalties could not be expected to affect their behaviour. 

Said differently, 76% of the offenders were “lacking at least one of the necessary conditions for making a rational response to punishments.  This group would be unable to make informed, systematic decisions about their crimes.  Furthermore, [the survey demonstrates that] 89% of those convicted of crimes involving death of the victim, 91% of sex offenders, and 88% of robbers “may lack the requirements necessary to make informed, rational judgements and to respond as intended to harsher punishment” (p. 305).   

Conclusion:  “The research suggests that the popular strategy of addressing crime with adjustments in the penal code is unlikely to provide substantial reductions in crime rates and that solutions to the… crime burden must involve a new emphasis on alternative deterrents.  The findings speak against more severe sentencing, not for emotional reasons, but because most current criminals do not have the information or mindset required to respond to these incentives for compliance” (p. 308).   For example, 89% of the most violent offenders were not thinking about the possibility of apprehension or the likely punishments associated with their crimes.

Reference: Anderson, David A. (2002). The Deterrence Hypothesis and Picking Pockets at the Pickpoket’s Hanging.  American Law and Economics Review, 4 (2), 293-313.

Back to top

The development of ‘new and better’ community sanctions is unlikely to reduce prison populations unless a serious commitment to reforming sentencing occurs as well.

England, in recent years, has begun to tackle its growing prison population, in part by attempting “to persuade sentencers, and the public, that community sentences are credible enough to use in place of some custodial sentences….” (p. 2).  This move was prompted by a growth in prison populations  that “could not be explained by increased crime or seriousness of offending” (p. 5).  Between 1998 and 2008, the number of people subject to community sanctions increased by 10%, but the prison population increased by a third.  Findings such as this lead many to question whether a growth in the use of community sanctions will necessarily lead to a reduction in imprisonment. 

Part of the problem with community sanctions is there are conditions attached to them, the violation of which can put the offender in custody.   With large numbers of conditions and increased enforcement of these conditions, breaches also increased dramatically during this period: there was a 470% increase in imprisonment for breach of non-custodial sentences between 1995 and 2009.   Not surprisingly, in this context, “all indications suggest that the growth in the use of community based sentences had a marginal displacement effect on custody” (p. 10).   Part of the reason that community based sentences may not reduce imprisonment is that imposing a community sanction on an offender may lead eventually to a prison sentence because many of the conditions of a community sanction have a high likelihood of being breached. 

Between 2004 and 2009, prison admissions for sentences of less than a year decreased by about 2,000 people (or 3%).  However, about 40,000 more offenders began a ‘suspended sentence order’ in 2009 than in 2004. In addition, the number of short prison sentences combined with a community based sentence increased dramatically.  At best, the promotion of community sentences beginning around 2005 may have stopped the increased use of short prison sentences, rather than causing a reduction in the use of prison. 

Conclusion:  This is not a new story.  The fact that a sanction is designed to keep offenders out of prison does not mean it will be successful in doing so, especially if the conditions are likely to be broken and offenders are likely to be imprisoned as a result.  “The data in this paper have told a clear story of the failure for those concerned with prison reduction…. It is not clear that a strong case for community sentence promotion as a mechanism [to address prison population size] exists…. There may be other valid reasons for supporting community sentences, such as a commitment to their potential rehabilitative ideal.  However, the evidence presented here suggests that the promotion and reform of community sentences will not fundamentally realize a longer-term vision for a significantly different, reduced custodial population” (p. 16).   It would appear, as one senior official suggested, that “What we have to think about is recalibrating our expectations about what sentencing is supposed to achieve” (p. 15). 

Reference: Mills, Helen. (2011)  Community Sentences: A Solution to Penal Excess? Centre for Crime and Justice Studies, London, England.

Back to top

The gap between the earnings of Black and White Americans increases after imprisonment. 

It is well established that those with criminal records have difficulty finding work after being incarcerated, and these effects may be greater for Blacks than Whites (Criminological Highlights 6(3)#2).  Similarly, youths who have been incarcerated rather than being given community sanctions are less likely to be in the workforce after their sentences have been served (Criminological Highlights 11(4)#4).  And the increase in wages that occurs as people gain more experience in the workforce is lower for those who have been imprisoned than for comparable people who have not been imprisoned (Criminological Highlights, 5(3)#7).  

This study looks at the impact of imprisonment on the gap between the earnings of Black and White offenders, all of whom were imprisoned.  Because being imprisoned is, in general, harmful to employment and income prospects, it is possible that the difference between the wages earned by Black and White Americans would decrease, since both groups would be, as a result of imprisonment, equally stigmatized.  Alternatively, there could be a multiplicative effect: being Black and being an ex-prisoner may be more harmful than would be expected if the two effects had independent negative impacts on wages. 

This study looked at the wages earned in legitimate work by Black and White residents of the State of Washington before and after they were incarcerated.  Specifically, wages (the average hourly wage when the person was working) for at least two years prior to incarceration in a Washington prison and for two years after release from prison were examined. Data were derived from state employment records.  If an offender was re-incarcerated, wages only to that point were examined. 

During the two years prior to incarceration, Blacks tended to earn less than Whites.  After incarceration, however, the gap between Black hourly wages and White hourly wages not only still existed; it increased over time. This effect – a widening gap between the hourly wages of Black and White workers after incarceration – occurred even when various other factors were controlled (e.g., employment history, education, offence type, length of incarceration, age).  For example, work history prior to incarceration has an impact on wages, but the difference between Black and White workers remains. Interestingly, however, the impact of work history on wages (for both Black and White workers) is less after incarceration than before, suggesting that a favourable work history does not help those who have been incarcerated as much as it does those who have not been imprisoned.  

Conclusion:  As people get older, their earnings tend to increase. However, “after release from prison, … the rate of growth is slight, and post-release wage growth is even slower for Blacks than for Whites.  Black wages increase, on average, about 21 percent slower than Whites each quarter after release” (p. 273).  The exact reasons for this could not be discerned from this study.  Clearly, however, incarceration – especially for Black Americans – “inhibits labour market prospects and other life-course transitions… [This study] points to the compound disadvantage faced by Black relative to White ex-inmates” (p. 273). 

Reference: Lyons, Christopher J. and Becky Pettit (2011). Compounded Disadvantage: Race, Incarceration, and Wage Growth.  Social Problems, 58 (2), 257-280. 

Back to top

Prosecutorial evidence against a person whose face looks trustworthy is not as persuasive as exactly the same evidence used against a person whose face is not seen as being trustworthy.

One traditional justification for deferring to judges and juries on questions of the credibility of witnesses in court is that judges and juries are able to observe the witness. The theory seems to be that by observing the demeanour of witnesses, people can determine whether they are telling truth. Evidence, however, suggests that ordinary people and even trained police officers are not very good at determining whether someone is telling the truth (Criminological Highlights 2(6)#8, 5(4)#5).  The focus on cues from people’s faces when they are giving evidence ignores, however, the possibility that some people simply look more trustworthy than others. In other words, observers of court hearings may be drawing inferences about the guilt of an accused simply from what he looks like, whether he testifies or not.

A group of university students rated each of 20 photographs of white males exhibiting neutral expressions.  The people in each photograph were rated on a number of dimensions including attractiveness, kindness, and trustworthiness. Two photographs were chosen – the one rated as being most trustworthy and the one rated as being the least trustworthy.  

A new set of university students was given two trial vignettes. Each vignette was accompanied by one of the photographs (counterbalanced across participants such that each vignette was equally likely to be paired with each of the photographs). There were four vignettes in total: two involving serious crimes (a robbery-murder and a sex-offence-murder) and two involving less serious crimes (a fraud and a car theft).  The prosecutor’s evidence was presented to the study participants one item at a time.  They were asked after each piece of evidence had been presented whether they were, at that point, convinced of the accused person’s guilt. 

There were no differences attributable to the rated trustworthiness of the person depicted in the photographs associated with the less serious crimes.  In other words, for the less serious crime, the same amount of evidence was necessary to convince people that the accused was guilty when the accused looked trustworthy as when he looked untrustworthy.  However, for the serious crimes (involving murder), more prosecutorial evidence was necessary to convince participants of the accused’s guilt when the accused looked trustworthy (5.9 pieces of evidence) than when the accused looked untrustworthy (4.2 pieces of evidence).  Said differently, it would appear that compared to someone who ‘looks trustworthy’, ordinary people are willing to infer guilt on less evidence when an accused has facial characteristics that, in our culture, make him look untrustworthy.

Conclusion: This study suggests that men whose faces make them look untrustworthy may be disadvantaged at their trials or perhaps when the police are looking for suspects for a crime. Quite independent of what the evidence is, people appear to draw inferences about what kind of person someone is from the physical characteristics of their faces. Participants in an experiment were more likely to conclude, on the basis of incomplete evidence against an accused, that he was guilty if he appeared to have an untrustworthy face than if he was perceived to be trustworthy, even though the objective evidence against the two accused was the same.  That this effect appeared only in the most serious cases (murders), suggests that in less serious crimes, people may attend more to the evidence rather than drawing inferences from the physical characteristics of the accused. 

Reference: Porter, Stephen, Leanne ten Brinke and Chantal  Gustaw (2010).  Dangerous Decisions: The Impact of First Impressions of Trustworthiness on the Evaluation of Legal Evidence and Defendant Culpability.  Psychology, Crime & Law, 16(6), 477-491.

Back to top

Publication Type

Volume Number


Issue Number