Investigating the effects of input method on student learning in an ITS
Aidan Bebbington
Department of Computer Science and Software Engineering
University of Canterbury
Abstract
Personalised one-to-one tutoring is known to be the most effective form of instruction. However, with limited resources, the situation in most educational settings falls far short of this ideal. Intelligent Tutoring Systems (ITSs) are computerised systems which have the potential to provide instruction which is pedagogically equivalent to personalised tutoring. Extensive research has been focused on improving the effectiveness of ITSs in a variety of ways. This report presents an evaluation of the effectiveness for learning of two different interaction styles: selection and typing.
Two interfaces for an ITS for thermodynamics called Thermo-Tutor are designed and developed. The designs are justified in terms of theory from psychology and cognitive science. Finally an evaluation is performed to measure the relative effectiveness of these interfaces and these results are analysed and discussed.
Thermo-Tutor is developed using ASPIRE; a general purpose authoring system for constraint-based ITSs. ASPIRE is used to reduce the development time and evaluate its effectiveness. The effectiveness of ASPIRE is also analysed and discussed.
The results related to the relative effectiveness of the two interfaces are inconclusive. However, results show that students did learn while using Thermo-Tutor and comments from students were generally positive. Therefore, development of Thermo-Tutor with ASPIRE was successful. This research also helped to identify some possible improvements for the future.
