OBJECTIVE: Prior research on the latent structure of psychopathology has indicated the presence of a multidimensional symptom organization and a higher-order factor. This study aimed to evaluate the reliability and validity of a developmentally informed model to disentangle unity and diversity. METHODS: We included 2512 children aged 6-14 years-old. Confirmatory factor analysis examined the validity and reliability of a bifactor model with one general dimension and ten dimensions of psychopathology: fear, somatic, distressful thoughts, low mood, low motivation/energy, inattention-hyperactivity, temper loss, aggression, noncompliance, and low concern for others, based on questions from Child Behavioral Checklist and Strengths and Difficulties Questionnaire. We analyzed the reliability of this model and competing models using omega statistics. External validity was examined using sex, executive functioning, and psychiatric diagnoses. RESULTS: The bifactor model demonstrated good overall fit. Reliability analyses indicated that most reliable variance in total scores was attributable to the general factor, whereas the reliability of specific dimensions was largely contingent on this overarching factor. Specific factors exhibited differential validity in terms of sex, executive function, and psychiatric diagnosis. CONCLUSIONS: Reliability indices primarily supported unity, whereas external validation favored diversity. These results encourage the extension of developmentally based models to disentangle unity from diversity.
Mendes et al. (Thu,) studied this question.