Learning analytics and evaluative mentoring to increase the students' performance in computer science

摘要

This work is devoted to presenting some strategies aimed at incrementing the students' commitment with their formative process and, at the same time, achieving a satisfactory evaluative procedure. The strategies described in this work have been applied in the subject "Computer Structure and Technology", which corresponds to the BSc in Computer Science syllabus at the University of Almería. Broadly speaking, the subject delves into internal aspects of computer operation, such as several devices found in the computer data path, memory organization, and the control unit. These issues are frequently very unpopular among students because they are not familiar with the deepest level of the operations and, therefore, they usually need to dedicate a huge amount of time and effort to understand them. For these reasons, a student's level of competency at the endo of the course is not usually very high and many students even drop out of the subject in its early stages. To overcome this situation, we have introduced two novelties in the learning methodology. On the one hand, we have taken advantage of possibilities that the Learning Management System of our university (Blackboard) offers using Learning Analytics to find out what activities are more relevant from an academic result point of view. On the other hand, we have proposed evaluative mentoring to personally deal with the students' necessities. Thus, the methodology combines the advantages of both, online and on-site activities. As an obtained result, we have studied the validity of the methodological strategies, detecting the issues that have more impact on the assessment results. This fact is very helpful to focus our attention on these issues in the following academic years. Moreover, we have demonstrated a significative improvement in students' competency levels with respect to previous academic years in the same subject and other subjects of the same course.