Peer collaboration in schools can be a locus for negotiation of power and cultural norms. Set in the context of ethnically and linguistically diverse urban schools, this study discusses the possibilities and limitations of utilizing video-mediated interviews to reveal multiple voices. Our research method is based on video-recorded peer-to-peer interactions wherein a small group of students engaged in mathematics tasks, as well as video-mediated interviews with individual students while watching the video of themselves engaging in group work. Using the theoretical framework of ideological and sociohistorical nature of voice and figured worlds, our analysis revealed the ideologies and sociohistorical norms that influenced the ways in which the students collaborate with their peers. Based on our analyses of the video-mediated interviews, we raise awareness of potential conflicts that could affect the generation and development of ideas through collaboration. We also discuss ways to use video-mediated interviews as a methodology to interrogate the norms underlying collaboration among students in school settings.