The duration of the delinquent career

View the most recent version.

The duration of a delinquent or criminal career is the length of time between the first and last incidents in the career. The date of the last incident – the "termination" of the career, or "desistance" from crime – is difficult or impossible to determine without tracking the subject to his or her death. This is usually not possible, although it has been done (Laub & Sampson, 2003). Various methods have been used to circumvent this problem (Kazemian, 2007). One approach is to define the date of desistance simply as the date of the last recorded incident during the period of observation (Farrington & Wikström, 1994). The problem with this approach is, of course, that it causes a downward bias in the estimate of career duration, since the durations of careers which continue past the end of the period of observation are underestimated. Another approach is to confine the analysis of the duration of careers to those which appear to be terminated. This raises the question of how long an offender must be inactive to be considered to have desisted. Various periods of qualifying inactivity have been used as indicators of desistance, ranging from 1 year to 5 years (Kyvsgaard, 2003). Obviously, the longer the period required as evidence of desistance, the less likely one is to misidentify a career as terminated. However, given a fixed and limited period of observation, the longer the period of inactivity which is required as evidence of desistance, the shorter the remaining period of activity which is available for analysis. For example, if the period of observation is 10 years, and 5 years of inactivity is required as evidence of desistance, then the longest possible career is only 5 years. Thus, the choice of the indicative period of inactivity represents a trade-off.

In the present study, the period of observation is 10 years: the data are truncated at the 18^th birthday for the cohort born in 1987 and at the 15^th birthday for those born in 1990. Some guidance as to a reasonable criterion period of inactivity was available from an analysis of the distribution of amounts of time which elapsed between all pairs of adjacent recorded incidents in all repeat offenders' careers. Ninety-three percent of these intervals were of less than 2 years.¹ Therefore, a career was defined as terminated, or completed, if there were no incidents during the last 2 years of observation. The result is that completed careers have a maximum length of 8 years less a day: from the 5^th birthday to the day before the 13^th birthday for the cohort born in 1990, and from the 8^th to a day before the 16^th birthday for the 1987 birth cohort. Careers with an incident during the last 2 years of observation were treated as possibly continuing, and therefore of unknown duration, and omitted from the analyses of duration.

Another issue which arises in the analysis of career duration is the treatment of offenders with only one recorded incident, who constitute the majority, or at least the modal category of offender, in most criminal career research using samples from the general population (versus populations of high-risk subjects). Some researchers omit these offenders from analyses of career duration, on the grounds that the concept of "duration" (and even perhaps of "career") does not apply to one-time offenders. Others include one-time offenders, assigning their careers a duration of 0 (e.g. Kyvsgaard, 2003). The latter approach is used in this study, on the grounds that this is numerically a very important group, making up 69% of offenders in the two cohorts. However, one-time and repeat offenders are distinguished in the analyses, so that one-time offenders do not skew the duration estimates.

The distinctions discussed above result in a threefold classification of offenders and their careers for the purpose of analysing career duration:

• One-time offenders with apparently completed careers (i.e. no recorded incidents in the last 2 years of observation), and therefore a duration of 0;
• Repeat offenders with apparently completed careers (i.e. no recorded incidents in the last 2 years of observation); and
• One-time and repeat offenders whose careers may not be complete, because they have recorded incidents in the last 2 years of observation, and whose career duration is therefore unknown.

The distribution of offenders into these three categories, by birth cohort and by sex, is shown in Table 2.

Forty-one percent of offenders began and apparently ended their careers of recorded crime at least 2 years before the end of the observation period: 26% of offenders born in 1990, desisting before their 13^th birthdays, and 48% of those born in 1987, desisting before their 16^th birthdays. About four-fifths of these desisters have only one recorded incident in their careers. The lower proportion of completed careers in the younger cohort reflects the fact that a greater proportion of that cohort began offending in the last 2 years of observation (i.e. aged 13 or 14 years), and are therefore by definition incapable of having completed careers. The distributions do not differ greatly by the sex of the offender, although girls are more likely than boys to have completed careers, especially completed careers comprising only one recorded incident.

Most of the offenders in these two birth cohorts had rather short delinquent careers. Excluding the offenders whose careers continued into the last 2 years of the observation period, and therefore may continue into the future, the mean length of offenders' careers was 0.26 years, or 3 months (Table 3). Over 80% of these offenders had a career that lasted only 1 day, either because the offender was involved in only 1 incident, or because all his or her incidents occurred on the same day.² Over 90% had careers lasting 1 year or less. Only 5% of these offenders with completed careers had careers lasting more than 2 years.³

---

Notes

1. Some of the longer intervals – which ranged all the way up to almost 10 years – may be the invalid results of false positive matching of incidents involving different offenders (see the Methodology section).
2. Multiple incidents on the same day may be due to the UCR2 counting rules for incidents; see note 3 in "Changes in the seriousness of offending over the delinquent career."
3. If the population is restricted to repeat offenders with completed careers, then the mean career length increases to 1.23 years, or 15 months. The percentage of offenders with careers of 1 year or less decreases to 57%, and those with careers of more than 2 years account for 23% of the group.