ENHANCING CONCEPTUAL UNDERSTANDING WITH DATA DRIVEN LABS

Abstract

Teaching introductory statistics with a data-driven curriculum presents many challenges for the instructor. One challenge is to provide students with opportunities to work with data in a realistic context. If not done carefully, students spend their time struggling to learn the software, not engaging with the data. Students might be able to follow step-by-step instructions to "see " how data analysis is done, but still fail to connect this to important concepts. We report on a project to create a set of data analysis activities that use Fathom to engage students in exercises that emphasize the challenges of statistical inference beginning in the very first week of the course; involve students with real data and real research questions; and require students to discover analysis procedures on their own. The resulting set of labs emphasizes simulation and randomization-based inference procedures while working in the context of real data.