QUALITY ESTIMATION OF LEARNING OBJECTS IN REPOSITORIES OF OPEN EDUCATIONAL RESOURCES BASED ON STUDENT INTERACTIONS

Open educational resources have emerged as one of the cornerstones of open education. One of the main barriers hampering their use and adoption is the lack of sustainable and effective quality control mechanisms in digital repositories. Evaluation strategies such as peer review have not been sufficiently scalable to keep up with the fast pace of open content creation by the user community. This study presents a new approach grounded on learning analytics in order to estimate the quality of learning objects based on the interactions that students have with them in open environments. For the study, 146291 sessions of student interactions with 256 learning objects distributed through an open repository were analyzed. A total of 11 student-learning object interactions were considered in the study. The quality of the resources was measured using the standard evaluation instrument LORI (Learning Object Review Instrument). To study the relationships between the student interactions with the learning objects and their quality as well as to build a predictive metric, linear regression analyses were used. The results show that there is a relationship between interactions and quality, and that it is possible to estimate with a moderate error the quality of the learning objects based on the interactions that students have with them. The results obtained point out that the proposed learning analytic can be used in open learning object repositories to automatically detect conflicting or low quality resources.