F09: Teaching Confounding

Milo Schield (Augsburg University)


For Doers and Learners. Confounding is a big new topic in the 2016 update to the GAISE guidelines. Teaching confounding is arguably the biggest change in introductory statistics content in 50 years. Share your experience, thoughts and concerns on these questions: 

  1. How can we teach confounding without multivariate regression?
  2. How can we teach students about Simpson's Paradox if they can't work problems?
  3. How can we teach confounding without making our students distrustful of statistics as a discipline?
  4. How do we deal with causation in observational studies?
  5. What are some techniques that will help students evaluate an association as evidence for causation?
  6. What's involved in offering Statistical Literacy?

Placecard: Teaching Confounding