Abstract:

In higher education, the effective prediction of the learning outcomes of the students during the semester is a key factor to allow lecturers to balance the evaluation tasks and the effort of the students in order to foster their learning and the acquisition of the right competencies before the final evaluation. The present paper describes the experiment driven during the four past courses with the students of the second year of an engineering degree at the University of Jaén.

The development of the specified competences for the degrees in higher education in Spain has increased the number of evaluation tasks during the semester from the previous pre-Bologna system. From that system, that involved only master classes and a final exam, we have evolved to a new paradigm based on a continuous evaluation, which supposes a higher number of short assessment tasks for theoretical and practical topics. Those tasks, that are aimed to measure the acquisition by the students of the learning outcomes set for the subject, sometimes could not suit properly the desired objectives, mainly in time and required effort. Therefore, to detect how the students perceive the difficulty of each evaluation task, they were surveyed after each task, but before they know their grade. Thus, each student can submit an estimation of the grade they will obtain after the lecturer revision.

After analyzing the results, comparing the estimations made by each student and their final grade, the students can be classified into three categories: pessimist, or those who thought that they would obtain a lower grade than the obtained one, realistic, or those whose predictions were close to their final grade, and optimistic, who estimated a better grade than the one they finally obtained.

Moreover, the degree of the difference between the actual grade and their estimation will be employed to determine if the learning outcomes for each category are appropriate. This degree will give information for each student about their special deficiencies (optimistic with low grades), and even, to detect those students with high learning capabilities (pessimistic with high grades). Furthermore, that factor could detect some problems in the requirement level of the evaluation task, e. g., all the students are realistic but the grades are medium-low which could eventually mean that the subject has a poor requirement level.
Therefore, in this paper, we present the experimental setup and the results obtained for the grade perception from the years under study.

Keywords:

Self-assessment, grade estimation, higher education.