This study aimed at testing the dimensionality of the Need for Cognition Scale (NCS-18), controlling the effect of negative items. Participants were 215 undergraduate students from João Pessoa (PB) with mean age of 20 years (SD = 3.79). They answered the NCS-18 and demographic questions. Three models were tested: M1 (conventional one-factor model), M2 (two-factor, with positive and negative items loading in separate factors), and M3 (one-factor model, controlling for the effect of negative items). Results showed that M2 and M3 models had better fit indexes compared to M1. Nevertheless, M3 also had better fit indexes than M2. In conclusion, evidence that the NCS-18 is more adequately represented by one-factor model was observed, especially when the effect of negative items is controlled.
for cognition; negative items; scale; validity