The overall goal of this project will be to create and validate a "learning map" (Stat-LM) for the content of undergraduate introductory statistics. This learning map will be a graphical representation of statistics concepts with connections among the concepts suggesting effective learning sequences. Use of the learning map will improve undergraduate learning by providing diagnostic information to instructors about students in their statistics courses, informing professional development for undergraduate statistics instructors, and modeling how critical prerequisites taught in high school connect to postsecondary learning expectations. In the first phase of the project, researchers will collaborate with statistics instructors from five institutions to create a learning map representing the broadest possible content set for introductory statistics material and the typical learning patterns of students in statistics courses taught in high school, community college, and university settings. In the second phase of the project, the research team will study the accuracy of the learning map using data collected from the Comprehensive Assessment of Outcomes in a First Statistics Course (CAOS), a confirmed valid and reliable assessment of student learning in statistics. This analysis will provide evidence of the learning map"s accuracy as a model of how students develop knowledge of introductory statistics.
The project will build on research conducted at the University of Kansas in the development and application of learning maps as tools to support instruction and assessment by addressing two specific aims. First, the Stat-LM will be developed by expert statistics educators, educational psychologists, and statistics instructors from five varied types of institutions to model the broadest possible set of topics taught at the university level, thereby assuring its relevance for high school and undergraduate courses. Expert reviewers will provide feedback on the nodes and connections in the map in order to refine the content validity of the Stat-LM as a model of student learning. Second, the analysis using quantitative data collected from the CAOS will consist of an alignment study of the test items on the CAOS to the Stat-LM, identifying the nodes in the Stat-LM that students must have mastered to answer each item correctly. Along with data from the CAOS found to be aligned to the map, the map specification will be analyzed with diagnostic classification (also called Bayesian network methods; e.g., Rupp, Templin, & Henson, 2010). Based on results from these analyses, the researchers will evaluate the Stat-LM with respect to the ordered connections between nodes and the validity of nodes themselves as latent constructs. The Stat-LM will become a "living document", subject to refinements throughout the proposed project and thereafter. As the goal of this project is to produce a tool for aiding statistics instruction, versions of the Stat-LM will be shared broadly for use within varied instructional settings.
