Can Computer-Based Assessment be Used for Motor Skills Learning? A Training Transfer Study

Abstract

A study was conducted to assess the effectiveness of a Virtual Environment-based intelligent tutor developed to train maneuvering skills on a remotely operated underwater vehicle (ROV) and to evaluate training transfer from the VE-enhanced training environment to maneuvering an actual ROV in an underwater environment. Thirty-two subjects were randomly assigned to one of four test groups: full training, simulation only, ROV only, or no training. Participants completed two pre-test trials, followed by training, and then two different of post-test tasks. Pre-test trials and training transfer trials were conducted in a test tank using an actual ROV. The full training group received computer-based verbal coaching, performance feedback, a VE-enhanced external, and performance based progression through a matrix of exercises organized by increasing task difficulty. Subjects in the simulation only and ROV only groups received an equal amount of unstructured practice time using the ROV simulation and the actual vehicle respectively. The No Training group performed only the pre-test and post-test trials. Comparison of pre-test and post-test performance was made across a full range of performance parameters. For most measures, no significant differences were observed between the different treatment groups. Where differences were observed, subjects with initial low ability in pre-test did not benefit from the full training system as much as the other forms of training. Transfer trials indicated that all forms of training showed significant positive transfer over that of the No Training group.