Computer task-based evaluation technique for measuring everyday risk-taking behavior

Abstract

Human risk-taking behavior is a major factor for accidents. Several techniques for quantifying human risk-taking tendency include questionnaire and observation methods. These techniques, however, have been questioned their validity and reliability. Our objective was to propose and evaluate a computer task-based evaluation technique for measuring everyday risk-taking tendency. In this technique, the users perform tracing a certain length of pathway, from start to goal, shown on the display by mouse. The system monitors the trajectory of the mouse cursor and detects the point of decisionmaking when users change their strategy from steering motion to ballistic motion as the mouse cursor approaches to the goal, yielding the level of risk-taking behavior represented by the Index of Difficulty (ID) at the location of strategy change. The results of experiment showed that IDs were highly correlated with probabilities of risk-taking behaviors obtained from 16 question items. © Springer-Verlag Berlin Heidelberg 2007.