Childhood Predictors and Adult Life Success of Adolescent Delinquency Abstainers

Natalie Mercer


by Natalie Mercer, David Farrington, Maria Ttofi, Loes Keijsers, Susan Branje and Wim Meeus
Utrecht University, Utrecht, The Netherlands
Cambridge University, Cambridge, UK

Background & Hypotheses
Research suggests that between 6-25% of adolescents abstain from delinquency (e.g., Barnes, Beaver, & Piquero, 2011; Chen & Adams, 2010; Moffitt, Caspi, Harrington & Milne, 2002). Understanding how or why some adolescents abstain from delinquency is helpful for understanding and preventing adolescent (minor) delinquency. Additionally, studying how or why some adolescents abstain may be most informative when abstainers are compared to different types of delinquent adolescents.
With this in mind, our study aimed to test three hypotheses regarding the nature of adolescent delinquency abstention: First, the linear hypothesis expected that the factors that predict adolescent delinquency abstention would be the inverse of factors known to predict serious delinquency. For example, if poor parent-child relationships predicted delinquency, the most serious delinquents would have the poorest parent-child relationships and abstainers would have the strongest parent child relationships, with the majority of adolescents falling somewhere in between abstainers and serious delinquents. Alternatively, the discrete group hypothesis expected that abstention would be the result of unique factors unrelated to the distinction between different groups of delinquents. For example, abstainers may be shy adolescents, socially withdrawn and excluded from their peer groups, whereas we have no such expectations for these characteristics to distinguish between who will be a serious delinquent compared to the majority of adolescents who experiment with delinquency and rule-breaking. Finally, we tested a third two-group hypothesis that the linear and discrete hypotheses may actually represent two different groups of abstainers.

This study made use of longitudinal, multi-informant data spanning from age 8 to age 48 from the Cambridge Study in Delinquent Development (100% male) to compare adolescent abstainers (n = 49), self-report delinquents (n = 239) and convicted delinquents (n = 117). First, we tested our linear and discrete hypotheses using binary logistic regressions to examine which individual and environmental childhood factors (ages 8-10) would predict adolescent delinquency abstention (ages 10-18) from self-reported delinquency. We also examined which childhood factors would predict convicted delinquency from self-reported delinquency. Second, to test our two-group hypothesis, we conducted a latent class analysis on the significant childhood predictors of abstention.

We found that there were five predictors of adolescent delinquency abstention. Consistent with the linear hypothesis, adolescent abstainers reported characteristics opposite to those of convicted delinquents (namely, abstainers were high on honesty, conformity and family income). However, unexpectedly, we also found that abstainers also shared some childhood characteristics with convicted delinquents (namely, low popularity and low school achievement). Consistent with our two-group hypothesis, a latent class analysis indicated that the mixed factors predicting abstention can be accounted for by two groups of abstainers: an adaptive group (n = 27) characterized by high honesty and a maladaptive group (n = 22) characterized by low popularity and low school achievement. To validate these two groups, we examined potential differences on adult outcomes. We found that at age 48, adaptive abstainers outperformed all other adolescents in general life success, as indicated by the obtainment of major developmental tasks (e.g., having satisfactory employment, accommodation, intimate relationship), whereas maladaptive abstainers only fared better than delinquent adolescents in terms of substance use and self-reported delinquency.

The results of this study question the validity of generalization from single behavioral outcomes (i.e., delinquent versus non-delinquent) to general development (i.e., unhealthy versus healthy). Overall, results suggest that abstainers can be either conformists with protective factors or adolescents who may be excluded from delinquent behavior due to impairments (i.e., social or cognitive) that require further attention.

Link to paper > Contact >
Back to Spotlights