Problem: Probability and statistics are vital tools in a wide variety of fields and professions, from astronomy to zoology, and its teaching can be improved substantially with resampling. Next to calculus, statistics is probably the most taught course on college campuses. It also may be one of the least successful--students strain to remember or crib from formulae that distract them from the key tasks of problem formulation and data analysis. Computer-intensive resampling (an innovative technique now widely accepted among statisticians but not yet widely taught) allows students to determine sampling distributions and solve problems empirically with simulations based on sample data, rather than memorizing formulae. Objectives: This project will provide workshops to assist instructors in teaching introductory statistics using a resampling approach. Resampling is the repeated drawing of simulated samples, often from the given data, to observe the behavior of some statistic or estimate of interest. Bootstrap, Monte Carlo, and resampling counterparts to permutation (randomization) methods also are included. Sometimes the term "computer-intensive methods" is used to refer to these methods generally. Activities: There will be an initial series of meetings with instructors who already teach resampling, or order to learn the issues that arise during resampling education. Next, workshops for interested statistics instructors will be held. Participants will be provided with the materials they need to return to their institutions and teach resampling. Follow-up questionnaires will be sent to participants to assess their experience. A web site will offer some of the same materials. Two follow-up workshops will be held in order to have in-person meetings with participants to discuss their experiences in teaching resampling methods--an "innovate, assess, adjust" cycle.
