CycleTalk: Data Driven Design of Support for Simulation Based Learning

Article type: Research Article

Authors: Rosé, Carolyn P. | Kumar, Rohit | Aleven, Vincent | Robinson, Allen | Wu, Chih

Affiliations: Carnegie Mellon University, 5000 Forbes Ave., Pittsburgh PA, 15213, USA. E-mail: [email protected], {rohitk,va0e,alr}@andrew.cmu.edu | US Naval Academy, 121 Blake Rd., Annapolis MD, 21402-5000, USA [email protected]

Abstract: In this article, we discuss the motivation for a novel style of tutorial dialogue system that emphasizes reflection in a simulation based exploratory learning environment called CyclePad, which was developed in order to offer students practice at design and optimization of thermodynamic cycles. We argue that while typical forms of exploration support offered in simulation based learning environments are meant to encourage the development of good exploration process skills, the instantiation of these typical forms of support in CyclePad are not sufficient, primarily because students don't choose to use them. We present a preliminary cognitive task analysis of design exploration tasks using CyclePad. Using this cognitive task analysis, we analyze data collected in two waves of formal data collection. Finally, we conclude with some system desiderata derived from our analysis as well as discussion of directions for continued experimental investigations related to tutor style.

Keywords: Tutorial dialogue, exploratory learning

Journal: International Journal of Artificial Intelligence in Education, vol. 16, no. 2, pp. 195-223, 2006