Age of Onset of Offending: Exploring the Validity of Self-Reports Across Time

Lila Kazemian, University of Cambridge
David P. Farrington, University of Cambridge

ABSTRACT
Much of the research dedicated to the study of the onset of offending has focused on the impact of early onset on other criminal career parameters (i.e. frequency, versatility, seriousness, and duration); other studies have explored the factors predicting early versus later onset. In this study, age of onset is used to assess the validity of self-reports across time. This research throws light on the benefits of prospective versus retrospective longitudinal data. Finally, it seeks to compare self-reported and official age of onset for different types of offenses.

This study uses data from the Cambridge Study in Deelinquent Development, a prospective longitudinal study of 411 London males. Teh offense categories were included in this paper: burglary, shoplifting, theft of vehicles, theft from vehicles, theft from machines, theft from work, assault, vandalism, drug use, and fraud. It is organized in three main sections. First, it presents the distribution of onset ages according to the ten offense types. Second, it compares the ages of onset (per offense type) reported by respondents at different periods of the life course, from ages 14 to 32. Finally, it contrasts these results with those of the onset of official offending (i.e., age at first conviction).

In general, results seem to show that the age of onset reported at later ages rarely concurs with prior self-reports. Indeed, with time, respondents generally tend to either overestimate the age at which they first committed an offense or simply deny ever engaging in such acts. This lack of stability observed in age of onset results over time supports the relevance of prospective longitudinal studies.

(Return to Program Resources)

Updated 05/20/2006