Despite the benefits of team diversity, individuals often choose to work with similar others. Online team formation systems have the potential to help people assemble diverse teams. Systems can connect people to collaborators outside their networks, and features can quantify and raise the salience of diversity to users as they search for prospective teammates. But if we build a feature indicating diversity into the tool, how will people react to it? Two experiments manipulating the presence or absence of a "diversity score" feature within a teammate recommender demonstrate that, when present, individuals avoid collaborators who would increase team diversity in favor of those who lower team diversity. These results have important practical implications. Though the increased access to diverse teammates provided by recommender systems may benefit diversity, designers are cautioned against creating features that raise the salience of diversity as this information may undermine diversity.

The emergence of team-assembly technologies has brought with it new challenges in designing and implementing socio-technical systems. Our understanding of how systems shape the team-assembly processes is still limited. How do systems enable users to find teammates? How do users make decisions when using these systems? And what factors explain the characteristics of the teams assembled? Building on existing literature from CSCW, computer science, and management science, we propose a taxonomy to characterize how systems influence team assembly. This taxonomy argues that two dimensions determine how systems shape team assembly: (i) users? agency, to what extent the system enables its users to exercise their agency, and (ii) users? participation, how many users the system allows to participate in the team-formation process. The intersection of these two dimensions manifest four types of teams enabled by systems: self-assembled teams, staffed teams, optimized teams, and augmented teams. We characterize each one of these types of teams, considering their qualities, advantages, and challenges. To contextualize these types of teams, we map the current literature of team-assembly systems using a scoping literature review. Lastly, we discuss ways through which these two dimensions alter users' behavior, team diversity, and team composition. This paper provides theoretical implications and research questions for future systems that reconfigure the organization of people into teams.

People and organizations are increasingly using online platforms to assemble teams. In response, HCI researchers have theorized frameworks and created systems to support team assembly. However, little is known about how users search for and choose teammates on these platforms. We conducted a field study where 530 participants used a team formation system to assemble project teams. We describe how users' traits and social networks influence their teammate searches, teammate choices, and team composition. Our results show that (a) what users initially search for differs from what they finally choose: initially they search for experts and sociable users, but they are ultimately more likely to choose their prior social connections as their teammates; (b) users' decisions lead to non-diverse and segregated teams, where most of the expertise and social capital are concentrated in a few teams. We discuss the implications of these results for designing team formation systems than promote users' agency.

The field of Computer Supported Cooperative Work (CSCW) has an enduring interest in studying and designing technologies that bring people together in partnerships, teams, crowds, communities, and other collectives. As the technologies enabling group formation have evolved, so too have the guiding questions pursued by CSCW scholars. This review outlines the trajectory of scholarship on group formation with an eye towards the most pressing future questions in this area. To understand how CSCW researchers have studied technology-enabled group formation, we systematically review articles published at CSCW from 1992 to 2018. Exploring more than 2,000 potentially relevant works, we identified 35 focused on technologies and group formation. Content coding and thematic analysis revealed four periods and six themes in the study of online group formation. These themes include: group composition, self-presentation, assembly mechanisms, recruitment, organizing structures, and group culture. Quo vadis? Based on our review, we offer recommendations for the next generation of CSCW scholarship seeking to understand and enable collectives joining together online.