Evaluating a Computerized Aid for Conducting a Cognitive Task Analysis
Abstract
Models that simulate cognitive processes have demonstrated considerable success in a variety of technical domains such as, intelligent tutoring, predicting the complexity of human system interactions, decision support and expert systems among others. Creating such models requires considerable skill in conducting a cognitive task analysis. The conduct of a cognitive task analysis is costly and labor intensive. As a result, a few computerized aids have been developed to assist in the process of conducting such analyses. However, none have been evaluated to determine how accurately and consistently users of such tools can create cognitive models. If such tools cannot demonstrate the creation of accurate models of cognitive tasks across users then such tools will be ineffective and unreliable. The research reported herein presents the results of experimentation, which focuses upon the evaluation of a computerized aid, specifically CAT-HCI (Cognitive Analysis Tool - Human Computer Interface), for the conduct of a detailed cognitive task analysis. A sample of users for a newly developed interface (tactical display) for the Army's Bradley A3 Fighting Vehicle were asked to model their knowledge of a routine task. Measures of the accuracy and of the consistency of the user generated models were recorded and analyzed. Accuracy measures the level of agreement between subject models and a baseline model, while consistency measures the level of agreement between subject models.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 2000
- Accession Number
- ADA375817
Entities
People
- Jeffrey R. Voight
Organizations
- University of Central Florida