Abstract
Objectives: Although cross-sectional investigations have found a bifactor structure of psychiatric comorbidity that includes a general psychopathology factor plus more specific factors, prospective evidence supporting the bifactor structure is still limited. We evaluated the structural stability (i.e., longitudinal invariance) of the bifactor model in comparison to an alternative structure, a correlated factors model without a general psychopathology factor. We also investigated the models' generalizability to change processes in psychopathology. Methods: The analyses were conducted on 10-year follow-up data from 5,001 respondents in the US National Comorbidity Survey. Invariance was evaluated through a series of nested invariance tests using confirmatory factor analysis, and the models' generalizability to change processes was investigated using change scores of disorder status. Results: The bifactor model and the correlated factors model exhibited an equal degree of strong structural stability over time. Only the bifactor model satisfactorily characterized the structure of temporal changes in psychopathology. Conclusions: The bifactor structure with a general psychopathology factor is stable over time and describes temporal changes in psychopathology. The findings support the notion that the general psychopathology factor describes a transdiagnostic etiology and may therefore provide a useful target for intervention and treatment.