Correctional Service Canada
www.csc-scc.gc.ca
Common menu bar links
FORUM on Corrections Research
The Life Span of Criminal Behaviour: What Do We Know?
In the study of recidivism, research on criminal careers is now considered the way of the future. A
career can be defined as a course or progress through life or as a way of making a living. Research
on criminal careers typically uses the first definition, meaning that a criminal career is a sequence
of offences during some portion of an individual's life. Criminal career research frames criminal
behaviour as something that develops during offenders' lives, not as an isolated incident in a
person's life at one point in time.
Rather than merely looking at whether a known offender commits another crime or not, criminal career
research uses a much broader concept of recidivism. It examines what proportion of a population gets
involved in criminal behaviour (prevalence), at what age the criminal behaviour begins (onset), at
what age it stops (desistance), how long the criminal career lasts (duration) and the number of
offences typically committed during the course of the career (frequency).
The study of criminal careers requires longitudinal data on offending. Most studies on criminal
careers focus on convictions until offenders reach their early 30s, which is the time period when
most convictions occur.
A recent article summarized three large-scale studies in the United Kingdom on criminal careers.
This article focuses on the results of one of the three: a prospective follow-up of 411 males from
age 8 to 32.
The subjects were from London, England, and most were born in 1953.
Of the 411 males in this study, about one in three (or 153) was convicted of a criminal offence at
some point before their early 30s. Adjusting this figure for the number of males at risk (that is,
not already incarcerated), the prevalence of convictions was 36.8%.
Breaches of conditional release, most motoring offences and many other minor non-indictable offences
(such as drunkenness or common assault) were not included in these results.
Vehicle theft (14.6%), burglary (14.1%) and assault (10.9%) were the most common offences
committed.
The study also found that the peak age for the prevalence of convictions was 17, when about 11% of
the males in the sample committed an offence and were subsequently convicted.
The number of offences committed per year peaked at age 17, with 16.8 offences per 100 males. These
results were based on offences, not convictions, because sometimes two offences lead to only one
conviction.
As well, in order to study separate offending incidents, only one offence (the most serious) was
counted on each day of offending. For example, sometimes a wounding incident results in convictions
for the wounding and for possessing a weapon. For this study, only the wounding offence (the most
serious) would have been counted. While this approach could lead to the under recording of separate
incidents committed on the same day, it was the best approach under the circumstances.
Some researchers(1) have proposed that there are two categories of offenders, frequents
and occasionals, and that members of both categories incur convictions at a constant (yet different)
rate during their criminal careers.
Using data from the London study, it was calculated that frequents incurred convictions at a rate of
1.14 per year (when not incarcerated) and occasionals at a rate of 0.41 per year. The average time
interval between convictions was 10.5 months for frequents and 29.3 months for occasionals. Nearly
half (43%) the recidivists in the study - that is, those with two or more convictions - were
estimated to be frequents.
The peak age of onset (the beginning) of a criminal career was 14 (4.6% of first convictions), with a
second peak at 17 (4.4%). The average age at first conviction for the 153 convicted males was 17.5
years old.
Rather than presenting the onset rate, which was based on all males in the sample who were still
alive, it might be better to present a hazard rate. This relates the number of first-time offenders
to the number of males without a criminal conviction. The hazard rate showed a clearer peak at age 17
because of the smaller number of men with no convictions at that age (compared with age 14).
Figure 1 shows the relationship between the age of onset of offending and the average number of
offences committed up to age 32. The average number of offences decreased as the age of onset
increased, from just over eight offences committed by those who started their criminal careers
between the ages of 10 and 13 to less than two offences by those first convicted between the ages of
21 and 32.
The age of desistance can only be determined with certainty when people die. Nevertheless, it was
calculated that at age 32, the males in this study had committed their last offence at an average age
of about 23.
Some researchers(2) examined the data in this study and investigated predictors of
persistence or desistance of offending after age 21. The best predictors of persistence were: rarely
spending leisure time with a father, heavy drinking and frequent unemployment during the teenage
years.
It seems that teenage offenders who were heavy drinkers and frequently unemployed were likely to
reoffend, since 17 of the 19 with these characteristics were convicted of a further offence.
Figure 2 shows the average length of criminal career (that is, the time interval between first and
last convictions) for the participants in this study. As shown, those who were first convicted at the
earliest age (10 to 13) were the most persistent offenders, with an average career length of about 10
years.
The average duration of criminal careers dropped sharply from those first convicted between the ages
of 14 and 16 (8.2 years) to those first convicted between 17 and 20 (2.7 years). This finding
suggests that males first convicted as juveniles were much more persistent offenders than those first
convicted as adults.
Overall, the criminal career (up to age 32) of the 153 convicted males lasted 5.8 years on average,
from age 17-and-a-half to about 23, and included 4.5 offences. More than a quarter (27.5%) of the
convicted males had a criminal career lasting more than 10 years, and more than a third (40.5%) had a
criminal career exceeding 5 years.
In general, offending in one age group continued into another. For example, of the 35 males convicted
between the ages 10 and 13, almost three quarters (71.4%) were convicted between the ages of 14 and
16, compared with more than 13% of those not convicted between 10 and 13.
Nearly three quarters (73.5%) of those convicted as juveniles (age 10 to 16) were also convicted
between the ages of 17 and 24, as opposed to less than a fifth (16%) of those not convicted as
juveniles. Overall, more than three quarters (78.3%) of those convicted as juveniles were also
convicted as adults, compared with less than a quarter (21%) of those not convicted as juveniles.
These figures show the considerable continuity in offending from juvenile to adult years.
Research on criminal careers also focuses on recidivism probabilities, which are often very high.
For example, of the 153 participants who committed one offence, more than two thirds (68%) went on to
commit another offence by age 32. Of 104 men who committed a second offence, almost three quarters
(71.2%) went on to commit a third. From the third offence on, the probability of recidivism was
almost always between 80% and 90%.
Specialization
About one third of the offenders (50 of 153) were convicted of violent offences (assault, robbery or
threatening behaviour). They committed an average of 1.7 violent offences each, but interestingly,
they also averaged 5.3 non-violent offences each. In fact, only 7 of the 50 violent offenders had no
convictions for non-violent offences.
A model was tested that assumed that violent offences occur at random
in criminal careers. The study data were applied to this model, and it
was concluded that offenders in this study did not specialize in violence.
Furthermore, violent offenders and non-violent but persistent offenders
were virtually identical in their childhood, adolescent and adult backgrounds.
Given this, violent offenders could essentially be considered as simply
frequent offenders, and efforts to reduce future violence should perhaps
be targeted at all frequent offenders rather than only at violent offenders.
Figure 1
Discussion
Research on criminal careers has many policy implications. For example, the distinction between prevalence and frequency directs attention to different strategies for reducing crime.
If prevalence is high and offending is thinly spread over the population, then strategies for reducing prevalence should be directed at the whole community (education or welfare programs). But, if prevalence is low and offending is narrowly concentrated, then the emphasis should be on policies targeting known offenders (rehabilitation or incapacitation).
Figure 2
It would also be very useful for correctional workers to know the likely course of criminal careers for different categories of offenders. Special attention could then be paid to offenders who are likely to reoffend quickly or who are likely to commit serious offences.
Similarly, although few studies have looked at this, it would be useful for judges to know the effects of different sentences on the future criminal careers of offenders. For example, the average length of prison sentence increases between age 17 and the mid-20s, but it may be that the rate of offending and the remaining length of criminal career decline during these years. Therefore, time served may eventually exceed the remaining criminal career length, meaning offenders are kept in prison beyond the point when they would have stopped offending anyway, which does the offender and society little good and wastes valuable prison space. Sentences could instead be tailored to the length of criminal career remaining.
Clearly, the criminal career approach is fundamental to much criminological research, theory and policy.
D.P Farrington, "Criminal Career Research in the United Kingdom,"British Journal of Criminology, 32, 4 (Autumn 1992): 521-536.
(1) A. Barnett, A. Blumstein and D.P. Farrington, "Probabilistic Models of Youthful Criminal Careers," Criminology, 25 (1987): 83-107.
(2) D.P. Farrington and J.D. Hawkins, "Predicting Participation, Early Onset and Later Persistence in Offically Recorded Offending," Criminal Behaviour and Mental Health, 1 (1991): 1-33.
