Pilot edition, Spring 2027. This version of the book is being tested in a few sections and changes during the semester.

Appendix F — Design notes for instructors

These notes explain why the book is built the way it is: how it lines up with the GAISE guidelines, what the research says about simulation-first teaching, and the reasons behind its less familiar choices. The Preface has the short version.

F.1 GAISE

The ASA’s Guidelines for Assessment and Instruction in Statistics Education (GAISE) College Report was revised in 2016, and a 2025 revision is being released in stages (GAISE College Report ASA Revision Committee, 2016; ASA, 2025). The 2025 revision carries forward the 2016 recommendations and adds three: communication, ethics and inclusion. The table follows its ten.

GAISE 2025 recommendation Where it shows up in this book
1. Teach statistics as a process of gleaning insights from data to inform decisions Every chapter opens with a real question and closes by answering it.
2. Emphasize communication, with attention to scope and limitations The four questions close every chapter. Two of them, “Who does it apply to?” and “Did one thing cause the other?”, are about scope.
3. Focus on conceptual understanding rather than algebra and formulas Simulation builds each idea. Formulas arrive as shortcuts for what was simulated. Normal calculations for individual values get one short chapter of their own.
4. Integrate real data with a context and purpose Real data only. Each StatLens link opens on the chapter’s own data.
5. Encourage multivariable thinking Confounding and study design in Unit 1; multiple regression in Unit 4.
6. Use software and apps to explore concepts and work with data StatLens for concepts; Jamovi (or R) for analyses.
7. Emphasize responsible and ethical conduct with data Light so far. Study design discusses sham surgery. More is planned for the summer revision.
8. Use evidence-based pedagogies that engage students Predict → Do → Explain activities that students run themselves in StatLens.
9. Use a variety of formative and summative assessments Guided Practice with solutions, exercises, the workbook, and MyOpenMath homework.
10. Implement inclusive strategies Free and open; runs in a browser; designed to WCAG 2.1 AA accessibility standards.

GAISE does not prescribe simulation-based inference or any topic order. Its learning outcomes are deliberately method-neutral. Two of them are central here: understanding the role random variation plays in inference, and recognizing that sampling variation is one reason a conclusion can be wrong. Simulation-first is one well-evidenced route to those outcomes, not a GAISE requirement. The 2016 report notes the growing use of simulation and resampling to teach the logic of inference.

F.2 The case for simulation first

Cobb (2007) argued that the standard course centers on normal approximations only because, historically, the simpler randomization logic was too expensive to compute. With computers, randomization can sit at the core, and it connects how data were produced directly to the logic of inference.

The evidence since is encouraging and modest.

  • Tintle et al. (2011) found that students in a randomization-based course learned significantly more about inference, with comparable learning on most other topics.
  • A follow-up (Tintle et al., 2012) found higher retention months later in the areas the curriculum emphasized, with no loss elsewhere.
  • A multi-institution study (Chance, Wong & Tintle, 2016) found similar gains for instructors new to the approach as for experienced ones.

These are encouraging results, not proof that simulation-first is better for every student. The fair summary is that students do at least as well overall and better on the concepts of inference.

F.3 Why the less familiar choices

The four questions. The ASA’s statement on p-values (Wasserstein & Lazar, 2016) says conclusions should not rest on whether a p-value crosses a threshold, and that a p-value does not measure the size or importance of an effect. “Could it be chance?” is only the first question for that reason. “How big is it?” asks for an estimate, which is the “new statistics” emphasis on effect sizes and intervals (Cumming, 2014).

Confidence intervals as the nulls you wouldn’t reject. Teaching the interval as the set of null values a test wouldn’t reject ties estimation and testing together. Every conclusion reports a size. The test-first order (randomization before bootstrap) is kept because the logic of a test is easier to simulate first.

Prediction only when students have a basis. In physics, students who predicted the outcome of a classroom demonstration understood it better than those who only watched, and watching alone did no better than no demonstration (Crouch et al., 2004). A prediction with no basis, though, is just a guess, so the book asks for one only when students can give a reason. Otherwise activities run Do → Notice → Explain. Active learning in general raises exam performance and lowers failure rates in STEM courses (Freeman et al., 2014).

The two “how many”s. Watkins, Bargagliotti and Franklin (2014) show how simulation can mislead. The noise in a simulated sampling distribution shrinks with both the number of simulations and the sample size, so students can misread simulation noise as a real effect of sample size. Running more simulations doesn’t remove the pattern. The book therefore names the two quantities separately. The number of simulations controls simulation error; the number of observations controls the evidence and the width of an interval.

The normal curve, split in two. The normal model for a statistic (standard errors, z, tail areas) is taught once, at the end of Unit 2, as a description of what the simulations showed. Normal calculations for individual values get their own short chapter at the start of Unit 3, where the normal curve is a model for data rather than for a statistic. Keeping the two uses apart stops the normal curve from being re-taught in every method chapter, and follows GAISE’s push away from calculation for its own sake.

One home per idea. The 68-95-99.7 rule, “estimate ± 2 SE”, and the one-proportion simulation are each taught in one place, and later chapters refer back to them. Early reviewers found the book repeated these, and repetition is what made simulation-first feel slow.

References

  • American Statistical Association (2025). Guidelines for Assessment and Instruction in Statistics College Report, 2025 revision, released in stages. https://amstat.quarto.pub/college-gaise/
  • Chance, B., Wong, J., & Tintle, N. (2016). Student performance in curricula centered on simulation-based inference: A preliminary report. Journal of Statistics Education, 24(3), 114–126. https://doi.org/10.1080/10691898.2016.1223529
  • Cobb, G. W. (2007). The introductory statistics course: A Ptolemaic curriculum? Technology Innovations in Statistics Education, 1(1). https://escholarship.org/uc/item/6hb3k0nz
  • Crouch, C. H., Fagen, A. P., Callan, J. P., & Mazur, E. (2004). Classroom demonstrations: Learning tools or entertainment? American Journal of Physics, 72(6), 835–838. https://doi.org/10.1119/1.1707018
  • Cumming, G. (2014). The new statistics: Why and how. Psychological Science, 25(1), 7–29. https://doi.org/10.1177/0956797613504966
  • Freeman, S., Eddy, S. L., McDonough, M., Smith, M. K., Okoroafor, N., Jordt, H., & Wenderoth, M. P. (2014). Active learning increases student performance in science, engineering, and mathematics. Proceedings of the National Academy of Sciences, 111(23), 8410–8415. https://doi.org/10.1073/pnas.1319030111
  • GAISE College Report ASA Revision Committee (2016). Guidelines for Assessment and Instruction in Statistics Education College Report 2016. American Statistical Association. https://www.amstat.org/education/gaise
  • Tintle, N., Topliff, K., VanderStoep, J., Holmes, V.-L., & Swanson, T. (2012). Retention of statistical concepts in a preliminary randomization-based introductory statistics curriculum. Statistics Education Research Journal, 11(1), 21–40.
  • Tintle, N., VanderStoep, J., Holmes, V.-L., Quisenberry, B., & Swanson, T. (2011). Development and assessment of a preliminary randomization-based introductory statistics curriculum. Journal of Statistics Education, 19(1). https://jse.amstat.org/v19n1/tintle.pdf
  • Wasserstein, R. L., & Lazar, N. A. (2016). The ASA statement on p-values: Context, process, and purpose. The American Statistician, 70(2), 129–133. https://doi.org/10.1080/00031305.2016.1154108
  • Watkins, A. E., Bargagliotti, A., & Franklin, C. (2014). Simulation of the sampling distribution of the mean can mislead. Journal of Statistics Education, 22(3). https://doi.org/10.1080/10691898.2014.11889716