This study shows how the use of various measures of eye movement can serve to portray dynamic decision-making (DDM) in a coherent fashion. We extracted eye movement metrics relative to 1) scanpath, 2) eye fixations, and 3) pupillary response, to characterize DDM during the process of risk assessment. Results from Experiment 1 revealed that incorrect classifications were associated with 1) less efficient information search, 2) difficulties in making sense of critical information, and 3) a low level of cognitive load. In Experiment 2, we used eye tracking to assess the impact on DDM of introducing a decision support system. The addition of a temporal-overview display seems to affect processing time in DDM as indexed by shorter scanpaths and fixations during classifications. These findings illustrate how event-based eye movement measures can reveal characteristics and limitations of the ongoing cognitive processing involved in DDM and also contribute to usability testing.