<p>The similarity serves as a pivotal force in establishing relationships. Innovation tendency, which reflects a scientist’s intrinsic value orientation, is a critical factor in understanding scientific behavior. Previous studies have predominantly focused on individual-level analyses, neglecting to explore this issue from a group perspective. To address this, this paper introduces a novel group-level metric—Innovation Similarity Tendency Index (IST-index)—based on the concept of "assortativity" from complex network theory, which quantifies the extent to which the innovation similarity tendency among scientists shapes the relationships within a group. The findings show an IST-index value of 0.62. This value falls within the range signifying a strong effect, indicating that the innovation similarity tendency plays a significant role in the formation of collaborative scientific networks. Furthermore, we construct different subgroups based on three dimensions: group size, structure, and influence, and conduct a detailed analysis of IST-index variations across these subgroups. This study offers valuable insights into the emergent spontaneous order within group behavior, providing practical understanding for the formation, decision-making, and self-organization processes within scientific teams.</p>