Scuba diving simulator: testing real time decision making in a feedback environment

Abstract

In this thesis, human decision making in dynamic environment is studied. A scuba diving simulator, which is converted to a game, is modeled for this experimental study. Different versions and types of the games are played by subjects and the results are studied by analysis of variance method. Latin Square and repeated measures experimental designs are used in statistical analysis. Most of the dynamic decision making literature analyzes decision making in event driven environments where subjects have plentiful of time before making decisions. However, in our scuba diving simulator, classified as a clock-driven simulator, subjects make decisions continuously throughout the simulation. Since decisions are made continuously, the effect of game speed (time pressure) on performance and on learning is investigated. It is found that game speed is effective on subjects’ performances in most of the performance measures. As an extreme case, games with pause option are presented. However, it is observed that performance in games with or without pause option does not differ significantly. It is a known fact that in general, delays deteriorate performance and learning. In different types of games, material and information delays are added to the structure to account for their effect on performance and learning. Both information and material delays are found to significantly affect the performance. However, performance differences between delay and no delay games are decreased with practice. Subjects playing games get insight about dynamics of these systems. So, from a similar point of view, subjects who have prior experience might perform better. Although, it is expected that scuba divers will perform better than non scuba divers, no statistically significant difference is found between these two groups. However, performances of subjects differ significantly from each other.