Predicting Insufficient Learning of a Complex Procedure.
Abstract
Thorough learning of complex procedures is required in the design, production, and purchase of software. This study examines the utility of three sources of information for predicting errors in such learning. Twenty-eight subjects were asked to learn thoroughly a complex order handling and invoicing system. The subjects were then tested by having them fill out the invoice associated with a present order. During this simulation they were allowed reference to necessary data files but were not allowed reference to any procedural descriptions. The subject's prior knowledge and the subject's own estimate of the thoroughness of his or her learning were both predictive of overall performance, while study time was not predictive. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 11, 1977
- Accession Number
- ADA055586
Entities
People
- Don Lyon
- John C. Thomas
Organizations
- IBM Thomas J. Watson Research Center