Learning Analytics and Interactions in Virtual Learning Environments. A Comparative Study of Typologies and their Relationship with Academic Performance

Abstract

Analysis of learning data (learning analytics) is a new research field with high growth potential. The main objective of Learning analytics is the analysis of data (interactions being the basic data unit) generated in virtual learning environments, in order to maximize the outcomes of the learning process; however, a consensus has not been reached yet on which interactions must
be measured and what is their influence on learning outcomes.
This research is grounded on the study of e-learning interaction typologies and their relationship with students? academic performance, by means of a comparative study between different interaction typologies (based on the agents involved, frequency of use and participation mode). The main conclusions are a) that classifications based on agents offer a better explanation of academic performance; and b) that each of the three typologies
are able to explain academic performance in terms of some of their components (student-teacher and student-student interactions, evaluating students interactions and active interactions, respectively), with the other components being nonrelevant.