Statistical Analysis of Specific Learning Results National University of Chimborazo

Abstract

The article deals with a topic of particular importance for university educational evaluation, focused on studying the quality of the Specific Learning Outcomes test using psychometric theory. With this intention, the work followed a non-experimental quantitative approach, with a descriptive scope, using the Classical Test Theory and the Item Response Theory. The results indicate that the tests have good quality in the design, construction, application, qualification and reporting phases; however, the verification phase is of poor quality, since piloting, calibration, and reagent selection are not performed. On the other hand, within the study of the items, quality problems are evident in difficulty and discrimination, while in reliability they showed consistency and precision in the scores obtained by the exam. Furthermore, the tests, as a whole, have a low power of discrimination between the different skill levels of the supporters. Finally, the results suggest that either of the two psychometric theories can be used to study the Specific Learning Outcome tests.