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

Preface

This preface is for instructors. Students: the Welcome page explains how to use the book.

Why this approach

The book follows the ASA’s GAISE College Report (2016) and its 2025 revision. That means conceptual understanding over formulas, real data with context and purpose, active learning, technology used to explore ideas, and conclusions that state their scope and limits. Simulation-first is one well-studied route to those goals: students in simulation-based courses do at least as well overall and better on inference concepts. Design notes has the details and references.

How the book is built

One idea and one habit. Students meet just two themes, introduced on the Welcome page:

  • the idea, “What if we did it again?”;
  • the habit of closing every analysis with four questions: Could it be chance? How big is it? Who does it apply to? Did one thing cause the other?

Everything below is how the book is written, and none of it is named for students.

Simulation first, then the model. Simulation builds each concept. The mathematical model (the normal curve, t, χ², F) then formalizes it for inference. Calculation for its own sake is kept to a minimum.

The shape of the course.

  • Unit 1 (Explore) is data and description.
  • Unit 2 (Simulate) teaches the logic of inference once. Randomization tests come first, then bootstrap intervals. The unit ends at the normal model, taught once, as a description of what the simulations already showed.
  • Unit 3 opens with a short chapter on the normal curve as a model for individual data values (heights, test scores), where the coursepack teaches it.
  • Units 3 and 4 then take one method per chapter, and each chapter has three beats:
    • Simulate: a brisk run on the chapter’s own data, giving both the test and the interval.
    • Model: the formula, as a shortcut for what was just simulated.
    • Compare: the two side by side, the conditions, and a case where the shortcut fails.
  • Unit 5 (Foundations) covers probability and the binomial, with simulation where it earns its place: probability as a long-run frequency, and the binomial as the distribution students already built in Unit 3.

Everything on the UWL syllabus is still here: the z, t, χ² and F procedures arrive in the Model beat.

Estimation alongside testing. A confidence interval is the set of null values a test wouldn’t reject. Every conclusion reports a size, not just a verdict.

Two “how many”s. The number of simulations controls simulation error, which is how much a simulated answer moves from run to run. The number of observations controls the evidence and the width of an interval. More simulations never make the evidence stronger. The book reports simulated values to about two decimals and compares a simulation with a formula by rerunning it.

Predict, then do. Activities ask for a prediction only when students have something to base it on: an earlier chapter, everyday experience, or the data on screen. Otherwise they do first, then notice and explain.

One home per idea. The 68-95-99.7 rule, “estimate ± 2 SE”, and the one-proportion simulation are each taught once. Later chapters refer back to them rather than re-teaching them.

A few conventions. Paired differences are first − second, in the order the data are listed, and every example says so. Conclusions describe a direction in words, never by its sign alone.

Tools

The chapters are code-free.

  • StatLens (learnlens.org/statlens) is the learning layer. Each StatLens box opens a free browser tool already loaded with the chapter’s data.
  • Jamovi is for producing analyses. The Tech Tutorials and Projects show each analysis in Jamovi.
  • R is an alternative: every Tech Tutorial and Project also has an R tab.

Students need nothing but a browser.

Acknowledgments

This book remixes two open textbooks from OpenIntro: Introduction to Modern Statistics by Mine Çetinkaya-Rundel and Johanna Hardin, and OpenIntro Statistics by David Diez, Mine Çetinkaya-Rundel and Christopher Barr. Both are licensed CC BY-SA, and so is this book. Its simulation-first approach owes much to Introduction to Statistical Investigations by Tintle et al. and to Statistics: Unlocking the Power of Data by the Lock family. StatLens was developed at the University of Wisconsin–La Crosse.