With the passage of No Child Left Behind in 2001, data collection and data-based decision making took center stage in many U.S. schools (Earl & Katz, 2002; Herman & Gribbons, 2001). This qualitative study furthers our understanding of how policy has translated into practice by providing an in-depth look at four middle school teams from two schools that used data as a tool for stimulating teac her learning. It seeks to answer questions about how team structures that were intended to support member learning actually shaped team function. In my observations of team meetings, interviews with team members, and review of team documents, a logical and reinforcing relationship between form and function emerged. One structure in particular emerged as important: type of data. Type of data seemed to signal team members about their team's functional purpose, their role as members of the team, the processes the team would use in their work, and what knowledge should be shared or constructed. The four teams included in this study represent two cases divided by data type. In one case--consisting of two teams from two different schools--teams chose to collect and examine standardized test data. I refer to this case as Single Source Data (SSD) Teams. In the other case--consisting of two teams from two different schools--teams chose to collect and examine a broader array of data that included standardized test data, grades, attendance, office discipline referrals, and a series of behavioral assessments. I refer to this case as Holistic Data (HD) Teams.