Zhu, Xi2013-09-102013-09-102011-07https://hdl.handle.net/11299/156548University of Minesota Ph.D. dissertation. July 2011. Major: Sociology. Advisor: Professor David H. Knoke. 1 computer file (PDF); x, 123 pages, appendices A-B.This research examines the structure, antecedents, and consequences of transactive memory system (TMS), a team's shared cognition of who knows what and the process of collectively encoding, storing, and retrieving knowledge. For the structure of TMS, I conceptually distinguish teams' cognitive and knowledge structures (i.e., cognitive accuracy, consensus, and knowledge distribution) from cognitive processes (i.e., coordinate and trust in encoding, storing, and retrieving knowledge), and develop a new method to measure the TMS structure. For antecedents of TMS, I examine how patterns of social interactions among team members (as captured by the structures of various social networks and individuals' positions in networks) affect cognitive accuracy and consensus of teams and individuals. For consequences of TMS, I examine the effects of cognitive accuracy and consensus on team performance and individual job burnout. Empirical analyses for this research are based on three waves of survey data collected from 26 multidisciplinary mental health care teams over a two-year research period. The results demonstrate that the structures of social networks are important predictors of team cognitive properties. Particularly, highly accurate and consensual TMS are found in teams with centralized and less dense task-help networks. Individuals' network positions have certain but limited implications for individual cognitive outcomes, which reinforces the idea that shared team cognition is fundamentally a team-level phenomenon. Additionally, I explain where cognitive inaccuracy comes from with a dyadic-level analysis. The results suggest that cognitive inaccuracy arises in dense task-help networks because team members tend to overstate others' expertise when receiving task-related assistance from the others. Examining consequences of TMS, I show that cognitive accuracy and consensus interact with team knowledge stock in affecting team performance, with accuracy and consensus having greater positive effects on performance in teams with higher knowledge stock. At the individual level, I find that cognitive accuracy alleviates job burnout for team members. But this effect is limited to one dimension of job burnout - burnout related to ineffectiveness. Together, the empirical results provide a strong support for the arguments that shared team cognition is constructed through social interactions in teams and that shared team cognition has positive effects on teams and individuals.en-USCognitive accuracy and consensusJob burnoutSocial interactionsSocial networksTeam performanceTransitive memory systemsThesis or Dissertation