Team communications provide a rich source of data on team or macro- cognition. Efforts to automate team communication analysis are easing the burden of manual analysis and are facilitating the investigation of a variety of dimensions of these data. Communications flow is one dimension that has been tapped and shown to be relevant to team performance. In this study macrocognitive processes are manually coded in the context of a Noncombatant Evacuation Exercise. In addition, communication flow patterns are extracted for a portion of the communication data and mapped onto the manually coded processes. The human tags of the remaining data are then predicted based solely on communication flow patterns with flow data predicting 73% of the variance in human-coded macrocognitive stages and 42% of the variance in the macrocognitive processes. These results can help inform macrocognitive theory and present promise for automatic tagging.