Training interventions designed to improve coordination and communication in the cockpit increasingly emphasize the teaching of specific coordination skills. However, there is little guidance in the literature regarding the manner by which these skills should be selected. This investigation compared a variety of task-importance indices used previously with individual tasks in predicting the overall importance of team tasks. All of the indices demonstrated relatively poor reliability. Composite indices, including one newly derived index, demonstrated greater validity. The results are discussed in terms of implications for future research and for team-training design.