Confirmatory factor analysis of the Uruguayan version of Trapnell's Smart Scale to measure perceived intellectual competence

This paper deals with the analysis of the items and valuation of reliability and construct validity of the four-item Smart scale developed by P. Trapnell (1994), by running exploratory and confirmatory factor analyses, in an attempt to confirm the unidimensional structure described in previous research. Exploratory results suggest a one-factor model which accounts for the 74% of the variance, with a reliability of 0,88. The items have satisfactory discrimination indices (> 0,62) and factor loadings are statistically significant (p < 0,001) and high (> 0,77). In order to perform the confirmatory analysis the participant group was split at random into two subgroups. The unidimensional model has been identified, parameters were adequately estimated and validated by running multi-group analyses. The scale proved to be an adequate self-report measure of perceived competence by summing the 4 items, with practically no skewness in the distribution of responses.