An Issue of Granularity: Decomposing Redesigned Courses on Different Levels of Details Fei Li, fl0030@unt.edu Serhiy Polyakov (corresponding author), sp0055@gmail.com Svetlana Barnes, sbarnes@lis.admin.unt.edu William Moen, wemoen@unt.edu Hong Xu, hx0008@unt.edu Texas Center for Digital Knowledge, University of North Texas, PO Box 311068, Denton, TX 76203-1068; Tel: (940) 565-3563 The THECB LOR Project The Texas Course Redesign Project initiated by the Texas Higher Education Coordinating Board (THECB) aims to develop and distribute instructional resources to improve student learning outcomes and lower the costs of higher education. The redesigned courses will be freely available to all public institutions of higher education in the State of Texas. To assist the THECB in reaching its goals, the Texas Center for Digital Knowledge (TxCDK) at the University of North Texas proposed the development of a learning object repository (LOR).The THECB LOR will facilitate the discovery, evaluation and reuse of not only the entire redesigned courses but also components of the course content. In Phase I, a TxCDK project team developed a proof-of-concept repository that used the content from a redesigned course, U.S. History I, to demonstrate the functionality and potential of a LOR for reusing and repurposing learning objects. The redesigned course content was decomposed into nearly 300 learning objects varying in levels of granularity. Phase II is a two-year effort that is addressing not only a next version of the LOR but also administration, workflow, and policies necessary for a production system. Granularity used in the LOR A key issue in the development of learning object repositories is how to decompose materials into pieces at different levels of granularity that are useful for different instructional needs. Granularity is a term used in instructional design to describe the size of a unit of learning (Wiley, 2000). As the level of granularity increases, the likelihood of a learning object being used as intended by the originator increases. However, smaller learning objects are less likely to be used as intended, and they could be reused in this capacity with little effort (Harvey, 2005). While the literature indicates that more granular learning objects have an increased potential for reuse (Harvey, 2005; Wiley, 2000), there is little empirical data about which levels of granularity are appropriate for different purposes. The decision regarding learning object granularity is a trade-off between the possible benefits of reuse and the expense of cataloging (Herridge Group, 2002; Willey, 2000). In the case of the THECB LOR, the focus is on providing the complete course as one learning object as well as providing discrete learning objects derived from the complete course. A major task in Phase I was making decisions about the levels of granularity for the derived learning objects. The U.S. History I course has a relatively traditional structure, and the project team used that structure to determine the levels of granularity. The contents were broken down and reorganized into five levels of granularity: Course, Unit, Lesson, Topic, and Free-Standing Learning Objects. Each Course consists of several Units. Each Unit includes multiple Lessons, each of which covers several Topics. The repository is also populated with discrete yet valuable