The evaluation of future teachers’ attitudes towards statistics is of interest due to the importance that statistical training has in current society for citizens, in general, and, specifically, due to its relationship with scientific literacy in teachers. As evaluations must be based on valid and reliable measurements, such work largely aims to contribute metric outcomes. The present work consisted of obtaining evidence of the consistency and validity of the scale of attitudes towards statistics (EAE) when administered within a group of 542 university students undertaking teacher training degrees. After describing the group under study, the construct was considered from a multivariate approach suitable for ordinal data, employing confirmatory factor analysis (ULSMV) and decision trees (CHAID and CART). Outcomes indicate that examined students did not have positive attitudes and considered their mastery of statistics to be insufficient for conducting educational research. On the other hand, students displayed interest in being up-to-date and able to discriminate fact from fiction. Following the evaluation of eleven measurement models and their associated fit indices, the present study concludes by present evidence of the convergent validity and consistency of a construct structure pertaining to five factors and 24 items. Additionally, the usefulness of EAE items in predicting determined criteria was evidenced (self-evaluation of performance and indicators of the importance attributed by future teachers to science and research in their profession).