出版社：Universidad de Chile. Facultad de Ciencias Sociales
摘要：Measurement in the educational field of academic achievement of university students is considered empirical and quantitative. Hence, the main purpose of such assessments is to control educational systems and evaluation based on objective criteria (Long, Wendt, & Dunne, 2011). The aim of this article was to develop an Item Bank for General Knowledge Test composed by 20 specific domains of knowledge. Considering that an effective construction of a test requires organization and systematization of activities, progress in six domains are presented. The sample was composed by 6,794 university students. 1,526 items from different domains were evaluated. To calibrate the items, a nonlinear exploratory factorial analysis was performed. Difficulty and discrimination indices were obtained according to the classical theory of tests and the item response theory, and reliability indices as well. It was observed that 68% of the items have moderate difficulty and 32% of them have high or low difficulty. Internal consistency of the instrument showed high reliability values, up to .70. Further studies are needed in order to expand the item sample, and review items that showed inadequate indexes on discrimination, difficulty and reliability. This instrument allows measuring academic performance on an interval scale level and reducing the misclassification of students.
关键词：General Knowledge Test, items bank, classical test theory.;Test de Conocimiento General, banco de ítems, teoría clásica de los test.
