Should Data Science Be Part of Introductory Statistics? (AMS Sectional West, Boise, Idaho)

Authors: Ji Y. Son, Claudia C. Sutter, James W. Stigler

Abstract: Introductory statistics is a ubiquitous course, taken at scale by students across disciplines, and it is often the last formal exposure many students have to statistical reasoning. Yet in many introductory statistics courses, there is a conspicuous absence of working directly with real data, and the connection of statistics to modeling. At the same time, data science has elevated the exploration and modeling of data as central practices valued across many professions. This raises an important question for educational institutions: does data science belong in introductory statistics? We argue that modeling data is not external to statistics, but rather an organizing principle that can bring coherence to the concepts typically taught in an introductory course. Using examples from the CourseKata project’s modeling-first curriculum, we demonstrate how students with diverse mathematical backgrounds engage in modeling by repeatedly making connections between familiar introductory statistics concepts (e.g., mean and standard deviation) and a central modeling idea (e.g., DATA = MODEL + ERROR). Results from a performance assessment administered across nine courses at five institutions provide evidence of which data science competencies students reliably develop and where persistent challenges remain. Introductory statistics may be an opportunity hidden in plain sight to scale data science practices broadly.

Room information: Interactive Learning Center Room 402