Research around problem solving in collaborative groups has made progress, but several conceptual and methodological issues remain. These issues include the appropriate choices of units of analysis; the ability of current theoretical sets of macrocognition codes to capture group dynamics; detection and identification of potentially emergent phenomena within groups; and the extent to which multiple dynamics are integral to understanding groups. Using data from a complex engineering challenge, we applied methods drawn from complex-systems analysis to offer insight into each of these issues, showing the need for multiple dimensions when studying group dynamics and highlighting methodological difficulties when dealing with emergent phenomena. We suggest future research to improve the understanding of the complex dynamics of collaborative groups.