Evaluation of learning objects through eye tracking

Abstract

A learning Object (LO) is any entity that can be used, reused or guided for learning supported by technology. In the literature there are evaluation methods that considers aspects of design and presentation of LO; however, these methods are not automated, since they are based on instruments and expert opinions. This article describes an evaluation technique of LOs with time data of eye fixations collected in real time by an eye-tracking device. It is considered that all graphic elements should provide information to the user, so that a measure based on entropy can be useful. An experiment was conducted to compare different LOs. The results show that entropy measure is useful for improving the design of LOs. Finally, the results are discussed in comparison to some aspects of evaluation used by experts.