Criminal offending and mental health problems often co-occur. This study examined competing models to understand bidirectional associations between crime and mental health problems over time among at-risk men in the U.S. It was hypothesized that there would be significant cross-lagged associations of offending and mental health problems in early adulthood. Longitudinal data were drawn from 206 at-risk men enrolled in the Oregon Youth Study. Random intercept cross-lagged models examined bidirectional associations between mental health problems (assessed with the 53-item Brief Symptom Inventory) and self-reported offending in early adulthood across 6 time points from ages 19-20 to 29-30 years. Sociodemographic factors, prior levels, and common risk factors (i.e., parents' criminality, mental health problems, and socioeconomic status, as well as child age, antisocial propensity, and internalizing behaviors) were controlled during analysis. A robust association between mental health problems and offending was found for the early adult period, especially for total and violence offenses. Findings did not support the hypothesized positive cross-lagged effects. Instead, two cross-lagged effects emerged indicating that offending was inversely related to subsequent mental health problems. Childhood antisocial propensity and internalizing behaviors emerged as important predictors of stable between-person level differences in offending and mental health problems, respectively, although results differed as a function of offense categories (total, violence, property, drug). Findings indicate that the relation between mental health problems and offending is unidirectional, temporal, and partially spurious. Preventing child internalizing behaviors and child antisocial behaviors holds promise for reducing early adult mental health problems and offending.
Keywords: Early adults; Longitudinal; Mental health; Offending; Random intercept Cross-lagged panel model.