We present results from a field experiment that tests the effects of varying gender and linguistic group composition on performance and on group-members’ perception that their voice is heard when completing complex collaborative work within a low scrutiny environment. We randomize individuals enrolled in a postgraduate course populated by mostly women and non-native English speakers into small teams within larger, exogenously assigned seminar groups. Groups are tasked with complex and deliberative research assignments over three months. Using administrative and survey data, we find that a higher share of women in seminar groups significantly benefits the academic performance of group members—an effect driven by a positive effect on female native English speakers — while a greater proportion of women in small teams improves non-native language speakers’ perception of being heard.