An ITS for medical classification problem-solving: Effects of tutoring and representations

  • Authors:

  • Rebecca Crowley;Elizabeth Legowski;Olga Medvedeva;Eugene Tseytlin;Ellen Roh;Drazen Jukic

  • Affiliations:

  • Center for Pathology Informatics, and Departments of Pathology and Dermatology, University of Pittsburgh School of Medicine, crowleyrs@msx.upmc.edu;Center for Pathology Informatics, and Departments of Pathology and Dermatology, University of Pittsburgh School of Medicine, crowleyrs@msx.upmc.edu;Center for Pathology Informatics, and Departments of Pathology and Dermatology, University of Pittsburgh School of Medicine, crowleyrs@msx.upmc.edu;Center for Pathology Informatics, and Departments of Pathology and Dermatology, University of Pittsburgh School of Medicine, crowleyrs@msx.upmc.edu;Center for Pathology Informatics, and Departments of Pathology and Dermatology, University of Pittsburgh School of Medicine, crowleyrs@msx.upmc.edu;Center for Pathology Informatics, and Departments of Pathology and Dermatology, University of Pittsburgh School of Medicine, crowleyrs@msx.upmc.edu

  • Venue:
  • Proceedings of the 2005 conference on Artificial Intelligence in Education: Supporting Learning through Intelligent and Socially Informed Technology
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

We report on the first evaluation of SlideTutor --a cognitive tutor in a medical diagnostic domain. The study was designed to evaluate both the overall impact of the tutoring system on diagnostic performance, and the effect of using two alternative interfaces to the system. The case-structured interface provides a case-centric view of the task, and the knowledge-structured interface provides a knowledge-centric view of the task. The study showed a strong effect of tutoring on performance in both multiple choice and case diagnosis tests for all students. Performance gains were maintained at one week. There were no significant differences between interface conditions on performance measures. Students in the knowledge-structured interface condition showed a significant improvement in correlating their certainty to their performance, which was not true of the case-structured condition. Ratings on a survey of user acceptance were significantly higher for the knowledge-structured interface.