Adopting a Randomization-Based Perspective in Your Introductory Statistics Course: Roadblocks, Risks, and Rewards

• Activity Number: 505
• Type: Topic Contributed
• Date/Time: Wednesday, August 3, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Education
• Author(s): Rachel Mary Dunwell*+ and Todd Swanson*+ and Alan Rossman*+ and Patti Frazer Lock*+ and Megan Spencer*+
• Companies: Hope College and Rhodes College and California Polytechnic State University and St. Lawrence University and Rhodes College
• Keywords: bootstrapping ; permutation testing ; resampling ; introductory statistics ; pedagogy ; randomization techniques

Abstract

In some introductory statistics courses, the emphasis placed on traditional normality-based hypothesis testing and confidence interval construction is increasingly giving way to randomization-based analogues. The introductory statistics courses at a growing number of institutions, including Rhodes College, St. Lawrence University, and Cal Poly - San Luis Obispo, now feature randomization-based methods to some degree, and sometimes to a very large degree. Using tools like R and Fathom, students in such courses quickly bypass formulas and tables, and rely instead on techniques that allow greater visualization and intuition to conduct statistical investigations. Panelists at this session will provide a brief overview of how (and to what extent) they have incorporated randomization-based methods into their introductory statistics courses, and what the results have been. After addressing some obvious concerns like "How will students react when all the p-values are different?" and "Can your students still do a basic t test?" and "How will my colleagues in the biology department react?", the panel will open the floor for discussion.