The research question revolves around understanding the impact of gender diversity, specifically the proportion of women, on the success of scientific teams and the underlying mechanisms driving this impact. While existing literature indicates a positive link between gender diversity and team success, measured through metrics like collective intelligence and innovation, the reasons behind this remain unclear. Most previous studies rely heavily on bibliometric data, focusing on the mere presence of women rather than their active participation and interactions within the team. This study aims to address this gap by employing relational data, providing insights into how women's roles and interactions influence team dynamics and knowledge creation. The study's importance lies in moving beyond simple demographic counts to understand the actual impact of women on team functioning, which is crucial for advancing both scientific knowledge and fostering a just and equitable academic environment.

The literature review reveals a paradigm shift in understanding the role of diversity in teams. Early studies often viewed gender diversity negatively, associating it with decreased effectiveness and increased conflict. However, more recent research emphasizes the importance of full integration and participation of diverse members for successful teamwork and innovation. The authors highlight the critical distinction between mere presence (tokenism) and genuine integration of women on teams. A meta-analysis from multiple disciplines confirms that holistic integration of diverse members contributes to innovation. Previous research shortcomings include a reliance on quantitative metrics and a lack of qualitative studies examining team processes, interactions, and power dynamics. The authors argue for a multi-pronged approach that includes qualitative and mixed-methods research, analyzing the team development process itself, not just the final outcomes, to accurately assess the influence of gender.

The study focuses on 12 interdisciplinary university scientific teams participating in an institutional team science program from 2015 to 2020. Data collection involved multiple methods, including participant observation, focus groups, interviews, surveys, and social network surveys at multiple time points. The social network surveys, the primary data source for this article, focused on team members' relationships (mentoring, advice-seeking, leadership, and collaboration). Social network analysis was employed to examine the patterns of interaction and collaboration within the teams. Indegree and outdegree measures, scaled by the number of respondents, were used to quantify the extent to which individuals received and gave advice, mentorship, and leadership within the team. Additionally, three combined measures—collaboration, social, and professional support—were created to further analyze interaction patterns. Field notes from participant observation provided qualitative context and informed the creation of classifications, such as 'senior woman,' which encompassed women in leadership positions regardless of formal title. Statistical analyses were conducted using RStudio, including t-tests and correlation analyses to test the hypotheses.

The study tested five hypotheses related to gender roles and interactions within the scientific teams. Hypothesis 1, which predicted gender differences in mentoring, student committee participation, advice-seeking, and leadership, was not fully supported. While some team-specific trends were observed, the data did not consistently reveal gender disparities in these areas. Hypothesis 2, predicting men's greater likelihood of being perceived as leaders, was not statistically significant. Although men were more often identified as leaders, the difference wasn't significant. Hypothesis 3, examining correlations between various network types (leadership, advice, mentoring, student committees), found that leadership and advice networks were strongly correlated, but mentoring did not differ significantly from leadership or advice-seeking networks. Hypothesis 4 investigated correlations between social and collaborative relationships, finding no significant gender differences, although both men and women showed intertwined relationships. Finally, Hypothesis 5 tested whether the presence of a senior woman impacted non-faculty members' social connections, finding no evidence to support this.

The study's findings challenge the initial assumption that scientific teams simply reproduce existing gender inequalities. The lack of support for several hypotheses suggests that team norms and established team dynamics, rather than gender alone, may play a dominant role in shaping interactions within these teams. The authors posit that the team science program itself, with its emphasis on team building, training, and support, may have mitigated traditional gendered patterns. The concept of ingroup formation, where the inclusion of underrepresented individuals leads to reduced prejudice, is presented as a potential explanation. The authors acknowledge that the absence of statistically significant gender differences does not imply perfect gender equality, but it suggests the potential of team science structures and interventions to promote more equitable and inclusive environments. The authors highlight the importance of modifying existing structures to create opportunities for women's advancement.

The study's main contribution lies in challenging the assumption that gender automatically determines roles and interactions within scientific teams, suggesting that team-level factors and interventions can play a significant role in shaping these dynamics. Future research should focus on how team science can change the culture of science to better support diverse members and promote more equitable environments. This includes studies that examine how support for diverse teams translates into broader culture change, investigating the impact of scientific interdisciplinary teams on access for marginalized groups, and focusing on team development processes rather than merely demographic composition.

The study acknowledges several limitations. Some teams were hesitant to participate, resulting in limited data. The Hawthorne effect—altered behavior due to participation in a study—is a potential concern. The definition of 'senior woman' may have been ambiguous and the interpretation of survey questions could vary among participants. The study did not directly measure team success, and the analysis was based on a single time point, preventing the examination of temporal changes in relationships.