The co-evolution of language and psychopathology: parsing within and between person effects across childhood

Background. Childhood mental health disorders are pervasive, costing the U.S. billions of dollars annually, and conferring enduring negative outcomes. However, there remains limited consensus on shared risk factors that could inform targeted intervention. Receptive language may be one such factor that contributes to the emergence and maintenance of psychopathology. However, little is known about how language and latent dimensions of psychopathology co-evolve across childhood.Methods. This study investigates the longitudinal associations between receptive vocabulary and psychopathology (internalizing; INT, externalizing; EXT, superordinate P) in a large, sociodemographically diverse sample of children from the Future Families and Child Wellbeing Study (N = 4,898). Data from the Child Behavior Checklist and Peabody Picture Vocabulary Test was collected when participants were aged 3, 5, and 9. Longitudinal measurement invariance of the bifactor model was first tested using structural equation modeling. Next, a random-intercepts cross-lagged panel model (RICLPM) was used to examine the temporal associations between receptive vocabulary and psychopathology across childhood. Results. The bifactor model met criteria for partial longitudinal metric invariance, corroborating dimensions of INT, EXT, and superordinate p across all childhood. The RICLPM revealed that much of the variance linking vocabulary to psychopathology dimensions was accounted for by between-person and within-person autoregressive effects, rather than within-person cross-lagged associations. At the between-person level, receptive vocabulary was negatively associated with P and INT, INT was negatively associated with EXT, and P was positively associated with EXT and INT. Conclusions. Stable between-person differences across childhood may account for the associations between receptive vocabulary and psychopathology.