The impact of racial and gender homophily was non-significant for either instrumental or expressive ties

Hmm. Interesting. Too small a sample and quite a nuanced study so I am somewhat skeptical even though the results are consistent with my expectations. From Homophily of Network Ties and Bonding and Bridging Social Capital in Computer-Mediated Distributed Teams by Y. Connie Yuan and Geri Gay.

To test our hypotheses on homophily and the impact of bonding and bridging ties on performance, we used a sample consisting of 32 advanced engineering design students assigned to work together on a complicated task. The purpose of this advanced class was to replicate what students would encounter in actual work settings and organizations (NASA). Therefore, we believe that the results of this study can be generalized beyond a college student population. However, the generalizability of our research is limited because the results were obtained from one single sample.

Counter to our predictions, the impact of racial and gender homophily was non-significant for either instrumental or expressive ties. However, homophily in social characteristics demonstrated a strong impact. Specifically, location was found to have a significant influence on the formation of both instrumental and expressive relations, and group assignment significantly influenced the formation of instrumental ties. These results support Bradner and Mark's (2002) earlier findings that location had a significant influence on social interactions in computer-mediated communications, and that the provision of different communication media—including rich media such as videoconferencing—to connect people from different locations did not shorten the "social distance" between them.

In many discussions in the US, in particular regarding reading, there is a great deal of focus on the importance of race. I have long been somewhat skeptical. People define themselves on an almost infinite number of variables: race, ethnicity, gender, SES, class, culture, language, religion, cohort, family structure, etc. and which elements are most important and to what degree, I believe, change over time and under different circumstances. Trying to pigeonhole people by race and gender is too confining and likely to lead to bad decision-making.