BoF 1-01: Designing a Confounder-Based Statistical Literacy Course

Milo Schield (Augsburg University)


A confounder-based statistical literacy course is different: less than a 30% overlap with a traditional
statistical inference course.  Students see more value in a confounder-based course than in an inference-based course.  Participants will discuss the statistical needs of students in various majors,  why confounding is more relevant than statistical inference for most students, why the denominator should be called the diabolical denominator, how students can work multivariate problems without a computer or calculator, and what are the necessary conditions for a confounder to nullify or reverse an association (the Cornfield conditions).