The youth self-report version of the Child and Adolescent Dispositions Scale (CADS) measures three orthogonal temperament dimensions that are hypothesized to dispose children and adolescents to develop psychological problems through transactions with the environment. The CADS was developed using exploratory and confirmatory factor analyses (CFA) of ratings of a pool of items from which synonyms and antonyms of DSM symptoms were excluded in two population-based samples. The criterion validity of the resulting three CADS dimensions of negative emotionality, prosociality, and daring has been well established in multiple samples. We examined the factor structure and criterion validity correlations of the CADS using data on 14.0–18.0 year olds in the English community based Neuroscience in Psychiatry 2400 Study. Using CFA, the three hypothesized CADS factors were identified, which were validated by their predicted cross-cutting correlations with measures of anxiety, depression, and conduct problems. Nonetheless, the fit of the CFA was not acceptable, likely because the expectation of simple structure in CFA was not met. The same data were re-analyzed using exploratory structural equation modeling (ESEM) to identify sources of poor fit. Two of the three fit indices suggested acceptable fit for the three-factor ESEM solution (Standardized Root Mean-Square Residual = 0.047; Root Mean-Square Error of Approximation = 0.067; Comparative Fit Index = 0.878). If replicated in other samples, these results suggest that some prosociality items with cross-loadings on more than one factor should be considered for elimination or replacement to improve the measurement of these dispositional dimensions.
