Real users, real results: examining the limitations of learning styles within AEH

Abstract

This paper examines the current state of AEH (adaptive educational hypermedia) research into explicit learning style modelling for user personalisation. It addresses the problem of non-naïve test subjects, who are often in user trials, thus contributing to experimental bias. Instead, the authors suggest using real people, i.e. users with a range of backgrounds and abilities, in order to gain a truer insight into evidence-based research. We report on a study carried out with No statistically significant differences were found between experimental groups, learning style preferences or learning environments. We discuss the significance of this, and then critically analyse the use of learning styles in relation to this study and also in the wider context.real users: around 80 children at a UK primary school. The study investigated sequential and global learning styles as a personalisation mechanism in an AEH system. The user trial involved matching and mismatching users and learning environments to see if learning improved. The AEH system used by the children was DEUS, a new e-learning platform that is conceptually similar to WHURLE, an AEH that also used learning styles as its user model. No statistically significant differences were found between experimental groups, learning style preferences or learning environments. We discuss the significance of this, and then critically analyse the use of learning styles in relation to this study and also in the wider context.