Assessing How Team Task Influences Team Assembly Through Network Analysis Emily Kaven 1 , Ilana Kaven 1 , Diego G´ omez-Zar´a 2(B ) , Leslie DeChurch 2 , and Noshir Contractor 2 1 Evanston Township High School (ETHS) District 202, Evanston, IL 60201, USA {eakaven,iakaven}@eths202.org 2 Northwestern University, Evanston, IL 60208, USA dgomezara@u.northwestern.edu, {dechurch,nosh}@northwestern.edu Abstract. What traits make users appealing as potential teammates? How do the traits that users seek out in teammates stay constant and differ as the task changes? Our study explores the social networks and skills involved in teammate selection. We performed a quasi-experimental study to analyze teammate choices for three tasks: launching a start-up, surviving in a jungle, and running an election campaign. We conducted our study in one graduate and two undergraduate classes, where students self-assembled into teams using a team recommender software. We ana- lyzed our results using Exponential Random Graph Models (ERGMs). Our results indicate that (a) among all three tasks, prior relationships were important, while (b) the importance of certain project skills varied across the three task types when choosing potential teammates. Keywords: Team formation · Task type · ERGM · Social networks 1 Introduction As projects become more complex and require knowledge from different fields, people organize into teams to solve challenging tasks. Organizations are encour- aging their members to self-assemble teams, enabling them to search and choose the most appropriate teammates [7, 26, 30]. Research has supported that these self-assembled teams can achieve high levels of satisfaction, cohesion, and per- formance [3, 21, 27], but studies have been inconclusive concerning the decision- making processes. Thus, several scholars have explored how characteristics, such as competence, similarity, and familiarity [2, 6, 10] influence decisions. While the literature on self-assembled teams is expanding among different research disciplines, little is known about how the task type—the set of shared goals that gets transformed into plans and strategies [15]—affects individuals’ teammate choices. Related literature has explored the relationships between the c  The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 R. M. Benito et al. (Eds.): COMPLEX NETWORKS 2020, SCI 944, pp. 322–334, 2021. https://doi.org/10.1007/978-3-030-65351-4_26