Validation of the five-factor model of the Attitude Scale towards Statistics in mexican psychology students

Abstract

The positive attitude towards statistics promotes the study and use of this fundamental discipline to analyze data, hence the relevance of evaluating the attitudinal level and developing strategies that promote a positive attitude. The Auzmendi's Scale of Attitude towards Statistics (EAE-25) is one of the most used instruments in Spanish language to assess this attitude; however, it has not been validated in Mexico, which is important for its use in this country and to provide evidence of validity. The objectives of the study were to test the originally proposed model, to verify its reliability, to describe its distribution, to compare the averages between both sexes, and to verify its concurrent validity. A non-probabilistic sampling was used in this instrumental study. A sample of 383 Mexican psychology students was analyzed (73.6 % women and 26.4 % men). The EAE-25 and a scale assessing attitude towards research were applied. Confirmatory factor analysis was used to test the Auzmendi's five-factor model. The parameter estimation was performed by unweighted least squares, using polychoric correlation matrix. The internal consistency was estimated by the McDonald's omega coefficient (ro), the convergent validity by the average variance extracted (AVE), and the discriminant validity by the hetero-trait-monotrait ratio of correlations (HTMT). The mean differences between both sexes were tested through the t-test and concurrent validity through Pearson's correlation coefficients. The model of five correlated factors (confidence, anxiety, pleasure, usefulness, and motivation) showed good fit (x2/df =1.79, GFI =. 98, AGFI =. 97, NFI =. 97, CFI =.99, RMSEA =. 04, and SRMR =. 07). The five factors showed reliability (ro >. 80 in confidence, anxiety, pleasure, and usefulness, and ro =. 72 in motivation) and discriminant validity (HTMT