Expertise, Motivation and Teaching in Learning Companion Systems

  • Authors:
  • Jorge Adolfo Ramirez Uresti;Benedict du Boulay

  • Affiliations:
  • Deparment of Computer Science, ITESM-CEM, Apdo. Postal 50, Mó/dulo de Servicio Postal, Atizapá/n, Edo. Mex., Mexico 52926. juresti@itesm.mx/ http://webdia.cem.itesm.mx/juresti/;School of Cognitive and Computing Sciences, University of Sussex, Falmer BRIGHTON BN1 9QH, UK. bend@cogs.susx.ac.uk

  • Venue:
  • International Journal of Artificial Intelligence in Education
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper describes work carried out to explore the role of a learning companion as a teachable student of the human student. A LCS for Binary Boolean Algebra has been developed to explore the hypothesis that a learning companion with less expertise than the human student would be beneficial if the student taught it. The system implemented two companions with different expertise and two types of motivational conditions. An empirical evaluation was conducted. Although significant differential learning gains between the experimental conditions were not observed, differences in learner behaviour between these conditions were. In particular students in the motivated condition with a weak companion taught it many more times than in the other experimental conditions and in general worked harder. Finally, the experiment also suggested that learning companions might be confusing for students if they try to resemble human behaviour, i.e. if they do not perform exactly as they are told.