This research examines the association between paternal incarceration and children’s delinquency. Prior research suggests an association, although omitted variable bias is an enduring issue.

To help address issues related to unobserved heterogeneity, we employ a method uncommonly used in criminological research. Rather than comparing the children of incarcerated fathers to respondents who have never had a father incarcerated, we exploit the longitudinal nature of the National Longitudinal Study of Adolescent Health to generate a strategic comparison group: respondents who will have a father incarcerated in the future. We also examine two types of delinquency, expressive and instrumental, to infer plausible mechanisms linking paternal incarceration and delinquency.

When using "futures" as comparison cases, results differ from much prior work and suggest a spurious association between paternal incarceration and instrumental delinquency (e.g., theft). Paternal incarceration retains a significant effect on expressive delinquency, which is partly mediated by reduced attachment to fathers.

The association between paternal incarceration and expressive (but not instrumental) crime supports Agnew’s strain theory and elements of control theory. Our comparison group also offers important advantages in terms of addressing unobserved heterogeneity, and we think this approach would prove useful for other topics in criminology.