Abstract

GOMES, Cristiano Mauro Assis  and  GJIKURIA, Jenny. Comparing the ESEM and CFA approaches to analyze the Big Five factors. Aval. psicol. [online]. 2017, vol.16, n.3, pp. 261-267. ISSN 1677-0471.  http://dx.doi.org/10.15689/ap.2017.1603.12118.

Marsh (2007) asserts that no self-reported questionnaire of the Big Five factors shows an acceptable degree of adjustment in confirmatory factor analyses (CFA). Based on CFA results, it can usually be inferred that exploratory factor analysis (EFA) solutions are incorrect, as well as concluded that the instruments evaluated by EFAs do not measure what they intend to measure. The present study investigates this assertion, evaluating the structural validity of the Personality Characteristics Inventory (PCI), a self-reported instrument of the Big Five factors. Only one EFA investigated the validity of PCI (Gomes, 2012). We re-analyzed the data of this study (716 elementary and middle school students from a Brazilian school). The solution of eight correlated factors was investigated through CFA and ESEM (exploratory structural equation modeling). The model was found to be unacceptable through CFA, showing good fit through ESEM. Marsh's assertion correctly predicted the results. Structural validity of PCI could only be concluded through ESEM.

Keywords : validity; confirmatory factor analysis; structural equation modeling; personality.

        · abstract in Portuguese | Spanish     · text in English     · English ( pdf )