In group decision making situations, group members may have or may express preferences for decision alternatives only indirectly or incompletely, based on goals and on information collected from the environment. For example, group members undertaking a route-planning task may express a preference only for route segments, such as traveling from home to the supermarket. The method developed in this research uses information collected on task structure and on group information seeking and (implied) preferences to predict group decisions on a dynamic decision making task. An advantage of the method is that it relies upon data collection methods that are either unobtrusive (e.g. clickstream logging) or minimally intrusive (i.e. recording of conversations) to inform predictions. The method is illustrated using data from the performance of a simulated emergency response task by experienced emergency response personnel. Potential use of the proposed method in predicting group decisions in certain dynamic decision making settings is discussed.