Deception has been an important problem in interactive groups, impeding effective group communication and group work, yet deception detection in such a context remains understudied. Extrapolated from the interpersonal deception theory (IDT) and group composition research in cooperative contexts, this research proposes that group factors, including diversity and familiarity, have influence on the performance of deception detection. The measurement of group performance was not limited to success, as previous deception studies did, but included efficiency as well because it is fundamental to the effectiveness of deception detection. An analysis of data collected from a real-world online community found that behavioral familiarity had a positive effect, and gender diversity had a negative effect, on group success in deception detection. In addition, behavioral familiarity had a negative effect and functional diversity had a positive effect on the group efficiency of deception detection. The findings not only extend IDT in several important ways but also suggest the need to distinguish between noncooperative and cooperative groups, an important theoretical implication for group composition research.