Ji Y. Son (California State University, Los Angeles), James W. Stigler (University of California, Los Angeles)
Abstract
Although the goal of modern statistics is modeling, the integration of statistical tests and procedures under a modeling framework is generally withheld from students until they get to more advanced courses, often in graduate school. In this workshop we share our experiences teaching introductory statistics, from the very beginning, as modeling. We organize our course around the core idea that DATA = MODEL + ERROR, and then help students practice the connections of this big idea across (1) concepts, (2) representations (including R code, visualizations, and mathematical notation), (3) and a wide variety of contexts (e.g., politics, sports, psychology) that make the introductory course coherent. We explore these ideas in the context of a modeling-centered curriculum (CourseKata.org) that has been implemented successfully with a broad range of students in California's public college systems.