The duration of the career and the age of onset

View the most recent version.

It has repeatedly been found that an early age of onset is associated with a longer (and more serious) criminal career. However, this association has been found to be weaker or nonexistent when only the delinquent career is examined (Kazemian & Farrington, 2006). Table 4 shows the relationship between the offender's age of onset and the likelihood of having a career which is completed during the period of observation (i.e. has no incidents in the last 2 years of observation). This table clearly shows that the reason why the overall proportions of completed careers are low is the large number of offenders whose first offence occurred during the last 2 years of the observation period. When these offenders are omitted from the calculations (Table 4, section b) the overall proportion of completed careers increases substantially from 41% to 73%.

Table 4 also shows that there is no association between the age of onset and the likelihood of having a career which is terminated during the period of observation. When the probability of the latter is regressed on age of onset separately for offenders in each birth cohort, and for all offenders together, the coefficients are small and non-significant, and in different directions. Thus, these data provide no evidence that offenders with earlier ages of onset are more likely to have careers which continue from childhood into adolescence (for the 1990 cohort), or from adolescence into adulthood (for the 1987 birth cohort). However, the fact that early onset offenders are as likely as late onset offenders to have careers which persist beyond the year 2003 suggests that early onset offenders have, on average, longer careers; and that, in general, career length is negatively associated with age of onset.

In summary, the available data suggest that the likelihood of a childhood-onset delinquent career continuing into adolescence, or of an adolescent-onset delinquent career continuing into later adolescence and possibly into adulthood, is unrelated to the age of onset. Given the truncation of the period of observation at 10 years, this means that the observed duration of the career is negatively related to the age of onset: early onset careers tend to be longer than late onset careers during the 10-year period of observation.