Reconceptualizing Desistance: An Examination of the Effect of Latent Adult Behaviors and Latent Adult Cognitions on Desistance from Crime Over Time

Research within the field of criminology on the “age-crime curve” demonstrates that the onset of offending begins around the age of 12, peeks around age 17, and begins to decline as individuals transition into adulthood. This general pattern of offending behaviors has received significant support from a wide range of studies. While the trends in offending are well documented, theories explaining the transition and subsequent declines in offending are still highly debated within criminology, with the primary foci centered around either external turning points (e.g., marriage, employment, and parenthood), or internal changes (e.g., cognitive transformations and identity shifts). While theories focused on external and internal changes have each received support, both perspectives – at least implicitly – suggest desistance is a much larger latent process. In other words, a single event (e.g., marriage) or even a cognitive change is likely one small part of a much more comprehensive process of unobservable change over time. Coupling perspectives on external and internal changes with scholarship on adulthood, I propose a new and enhanced method to capturing the latent movement into adulthood via markers of adult status, and examine how these latent markers relate to changes in offending as a process that unfolds over time. Drawing on 11 waves of data and around 1,350 adjudicated youth from the Pathways to Desistance dataset, I use Item Response Theory modeling strategies to construct measures for Latent Adult Behaviors (LAB) and Latent Adult Cognitions (LAC). Theoretically, LAB operates as a stake in conformity which strengthens one’s bond to society through adult behaviors. Similarly, LAC functions as a conventional commitment to adult-like cognitions, thus strengthening one’s social bond and suppressing criminal offending. Results of a series of mixed-effects models reveal that between-person differences and within-person changes in LAB are associated with increases in offending over time. On the other hand, between-person differences and within-person changes in LAC are associated with decreases in offending over time. Interaction terms suggest that the harmful influence of higher levels of between-person LAB are reduced with increases in within-person changes in both LAB and LAC