Items' difficulty level determination based on a Statics test with parameters variation



Título de la revista

ISSN de la revista

Título del volumen


Institute of Electrical and Electronics Engineers Inc.


Looking for the improvement of the students' learning, the engineering school of EAFIT University is currently developing an evaluation and training system. In this system, each student has the option of doing 'dynamic' exercises in a specific field of a subject. Whenever a student uses the system, an exercise is generated with different parameters and values. The mentioned system (Evaluation System) has been used in the last semesters, achieving the improvement in the students' comprehension and learning level of the Statics subject. The system under development, allows the students to train themselves in different topics of the subject and, at the same time, it allows the teachers to evaluate the learning process. Based on the fact that in the course assessments, an exam with different parameters and values is presented every time a student accesses the system, it is possible that some students present tests with different complexity levels. Therefore, the test could be considered inequitable for some people. This paper presents an analysis of the difficulty level of a test generated with the system. The test was applied to two different groups of students which are taking the course in the 2014-1 semester. The first group took a test where the items' values and parameters were not changed; and the second one, took a test where the items' values and parameters change for each student. Based on the obtained results, a statistical study is made which intends to determine the difficulty and discrimination level of each of the test items, both for the dynamics parameters test and the fixed parameters test in order to finally determine how much varies the items' difficulty due to the parameters variation. This will help to generate more equitable tests in the future for the assessment of a group of students that are taking the course. © 2014 IEEE.


Palabras clave