Abstract:

The main objective of the active learning methodologies is to directly involve students in the teaching and learning process. Students must give up their traditional passive role and become a key actor in the generation of knowledge. This change carries as a benefit an improvement in their own learning process, since it demands learning by discovery and being able to analyse, apply and transfer the knowledge. These high level cognitive process help the student in the in-depth understanding of the contents, in the retention and relationships between the involved concepts and the development of reflection and decision-making skills. This alternative teaching-learning approach must come along with changes in the assessment model. Traditionally, the evaluation process is responsibility of the professor, either the preparation of the tests and their subsequent correction. The workload involved in these evaluation activities often restricts the number of tests that are performed, limiting the feedback provided to the student. This restriction confronts the active learning methodologies principles, in which, ideally, the students should be able to identify if their decisions are correct and if they have reached the educational aims expected in the different activities carry out.

In this paper, we describe a teaching experience based in active methodologies, which is being developed in a subject of the Degree in Computer Engineering. In particular, the focus is on the proposal of integrate an automatic test generation system. This system gives the students support in the evaluation process, in two ways:
1. in the generation of the tests they create for their classmates about the contents of the subject, which conveys the relationship and integration of the topics studied;
2. in obtaining feedback from the resolution of the tests produced by their peers and professors throughout the process.

The work describes the experience with the active methodology addressed, in which students work in teams to create educational resources which are used to present to their classmates certain topics studied and/or researched. As part of this process students are also responsible for creating the MCQ (Multiple Choice Questions) test on that contents, which will be used to provide immediate feedback, and in this way, enable the identification of the level of understanding by their peers. The MCQ automatic test generation system is based on semantic and service technologies. This system assists in the process of preparation of tests and dynamically generates the distractors (the alternatives to the correct answer) required for each of the proposed evaluation items. The combined use of semantic similarity algorithms and different heuristics allowed the configuration of the precision of these distractors and, therefore, the complexity of the test.

This work describes the system developed for the generation of tests and how its integration has been planned within the framework of the active experience to attend to the aspects of the assessment.

Keywords:

Automatic assessment, active learning methodologies, feedback, MCQ, semantic technologies.