When decision makers work collaboratively, their combined performance can exceed the performance of even the best group member working alone. However, despite the potential for performance gains when working in a group, collaborating individuals often show poor coordination and losses of motivation, and are generally inefficient at combining their resources and effort – all contributing to suboptimal group performance. The present experiment aimed to investigate how cognitive self-monitoring, or metacognition, changes under collaborative conditions. 54 participants (38 female, 16 male) formed 27, 2-person groups, and performed a gauge-aggregation signal detection task both individually and collaboratively. Measures of Type-1 (signal detection sensitivity) and Type-2 (metacognitive sensitivity) performance were calculated from individual and group confidence ratings. Hierarchical Bayesian parameter estimates suggested that when working collaboratively, groups showed no collaborative gains in task performance, but did trend towards outperforming the more metacognitively sensitive members. Similarly, groups showed no collaborative gains in metacognitive efficiency, and in fact trended toward metacognitive losses. Despite the potential for collaborative gains and increased metacognitive awareness, groups likely failed to show any benefits of teamwork due to poor integration of information, and suboptimal weighting of group member contributions.