Inventing the Digital Dashboard for Learning Diana Marie Bajzek, Office of Technology for Education, Carnegie Mellon University, USA db33@andrew.cmu.edu William E. Brown, PhD., Dept. of Biological Sciences, Carnegie Mellon University, USA wb02@andrew.cmu.edu Marsha Lovett, PhD., Eberly Center for Teaching Excellence, Carnegie Mellon Univ., USA lovett@cmu.edu Gordon S. Rule, PhD., Dept. of Biological Sciences, Carnegie Mellon University, USA rule@andrew.cmu.edu Abstract: Carnegie Mellon is creating a Digital Dashboard for Learning (DDL) to enable faculty and students to tune their immediate actions (lectures, assignments, study) based on rapid feedback on student progress in learning. The DDL will provide faculty with a timely overview of student progress in blended learning environments supported by the Open Learning Initiative, Carnegie Mellon’s project to develop high quality online courses and course materials. The DDL will provide access to details of learning measurements tied to course topics and “drill down” links to original presentation materials. It will enable faculty to aggressively adapt lectures and assignments to address weaknesses in measured student performance. This paper describes the requirements for the DDL, design and implementation issues, such as modifications in associated learning objects, tagging of student interactions, data collection and analysis, and the form and content of data presentation to the faculty and students. The Challenges There are many challenges shared by today’s learning environments and as the size of the classes increase or the frequency of direct contact with the student decreases, the magnitude of these challenges increases. First, there is great variability in the students of today. Their background knowledge, relevant skills, and future goals make it difficult for faculty to address students’ diverse needs (Felder & Brent, 1996; Fink, 2003). Students arrive with preconceived views of the topics and a false sense of security in having heard many of the topics at least mentioned in previous, lower level courses. Second, the larger one’s class, the harder (and more costly) it is to employ the best teaching practices that foster deep learning, e.g., personalized instruction, rich and timely feedback, and interactive learning environments (CFE, 2003; NRC, 2004). Third, although the conceptual structure of knowledge is clear to experts, it is not to novices. The array of new ideas and unfamiliar terminology in introductory courses tends to overwhelm students into memorizing a set of isolated facts without understanding the underlying common principles (Chi, 2005; diSessa, 1993). Finally, in many curricula, concepts are introduced initially in basic form and then applied and relied upon in multiple contexts without instruction that scaffolds these extensions of the concept to other contexts. Students who tend to compartmentalize what they learn not only become confused about the transfer of concepts to other contexts, but also miss opportunities to connect their knowledge and generalize their understanding. From research on how students learn, two well-supported principles have emerged that can be leveraged to address these challenges. First, students’ learning improves and their understanding deepens when they are given timely and targeted feedback on their work (Butler & Winne, 1995; Corbett & Anderson, 2001; NRC, 2001, 2004). By “feedback” we refer to corrections, suggestions, cues , and explanations that are tailored to the individual’s current performance and that encourage revision and refinement. Second, students benefit from an explicit conceptual framework that organizes the material they are learning. Effective instruction will to make that framework explicit and salient, and students need to practice making connections between related ideas in the framework (Eylon & Reif, 1984). We believe strategies that employ these principles in a less studied area of research, namely how faculty can best use information on their students’ progress to effectively adapt their teaching, offers an opportunity for addressing these challenges listed above. Equally, we believe that digital learning environments (web-based course materials, simulations, virtual labs, learning objects of many kinds) provide an ideal context for providing immediate feedback, supplying and continually reinforcing conceptual frameworks, and, most importantly for this paper, giving faculty the data they need to adapt their instruction to the learning needs of particular classes and particular students, and provide students with better tools to monitor their own learning progress.