The core self-evaluations personality construct : testing a bifactor model

The core self-evaluations personality construct (CSE) has been defined as a meta-trait comprising the subdomains of self-esteem (SE), generalised self-efficacy (GSE), neuroticism (N), and internal locus of control (LOC). The conceptualisation and operationalisation of CSE as a unitary construct has been criticised in its failure to account for the distinctiveness of the subdomains. To address this concern, our study tested a bifactor model, enabling the specific factors (SE, GSE, N, and LOC) to be empirically represented, independently of the general factor (CSE). Adult participants (N = 331; 65% community members, 35% undergraduate students) completed an online survey with established Likert scales of the relevant constructs. Confirmatory factor analysis revealed that the bifactor model provided the best fit for the data, compared to traditional representations of CSE as a single-factor or higher-order model. Each specific factor explained significant variance beyond its expression in the general factor. Structural equation modelling showed that the bifactor model was also superior to the higher-order model in predicting life satisfaction, a known correlate of CSE. The specific factors SE and LOC (but not GSE or N) contributed additional predictive utility beyond the general factor. Results of this study provide support for the distinctiveness and discriminant validity of the CSE subdomains.