Interactive Apps
Interactive Apps
A set of free, browser-based Shiny apps that accompany the book. Each one is referenced from a Try It Online callout in the chapter that introduces the underlying idea. No account or installation is required.
If an app’s link ever stops working, this is the canonical page to check — every chapter callout points here, so updates land in one place.
Chapter 1 — Why Statistics Matters Now
Variable Type Explorer
Classify variables from a hospital dataset as nominal, ordinal, discrete, or continuous. The “What Goes Wrong” tab shows what happens when you apply the wrong summary statistic; the “Numbers Trap” tab tests whether you can spot variables that contain digits but are not numerical.
Chapter 2 — Asking Good Questions: Research Design
Sampling Explorer
Compare simple random, stratified, cluster, and systematic sampling side by side. Draw repeated samples and watch how tightly the estimates cluster across methods. The same app supports the sample-size and margin-of-error sliders used in Chapter 7.
Total Survey Error Explorer
A clickable taxonomy of every survey error type, with definitions, practitioner examples, and notes on which errors larger samples actually fix.
Chapter 3 — Summarizing Data with Numbers
Datasaurus Explorer
Twelve datasets with identical means, standard deviations, and correlations — yet wildly different shapes. The visual argument for why summaries alone are not enough.
Chapter 4 — Summarizing Data with Pictures
Correlation Game
See a scatter plot, estimate \(r\), then reveal the answer. Builds visual intuition for what different correlation strengths actually look like.
The Datasaurus Explorer (Chapter 3) is also reused in Chapter 4 to make the case that pictures reveal what tables hide.
Chapter 5 — Probability
Probability Rules Sandbox
Set \(P(A)\), \(P(B)\), and \(P(A \cap B)\) with sliders and watch every probability rule update live on a Venn diagram, a unit-square view, and a tree diagram. Force-toggles for independence and mutual exclusivity.
Base Rate Lab
The COVID-test base-rate problem made visceral. Drag the prevalence slider and watch the positive predictive value collapse even though the test never changes.
Binomial Explorer
Set \(n\) and \(p\) with sliders and watch the binomial PMF and CDF update. Overlay the normal approximation and see when it succeeds and when it fails.
Chapter 6 — Normal Distribution and the Central Limit Theorem
Normal Curve Explorer
Set \(\mu\) and \(\sigma\), pick a region (less than \(a\), between \(a\) and \(b\), within \(\pm k\) standard deviations), and watch the shaded area and probability update live.
Distribution Explorer
The Central Limit Theorem in action. Pick a skewed population, set the sample size, and watch the sampling distribution of the mean transform from skewed to normal as \(n\) grows.
Chapter 7 — Confidence Intervals
Confidence Interval Simulator
Generate hundreds of confidence intervals from random samples and watch the capture rate. Switch between 90%, 95%, and 99% confidence and see the miss rate change.
Launch the Confidence Interval Simulator
The Sampling Explorer (Chapter 2) is reused here for sample-size and margin-of-error sliders.
Chapter 8 — Hypothesis Testing
Hypothesis Testing Playground
Set \(\alpha\), sample size, and effect size, then watch Type I error, Type II error, and power respond. Run repeated experiments under a true null and see how often you reject.
P-Hacking Simulator
A dataset with no real relationships and a menu of analysis choices. See how quickly you can reach \(p < 0.05\) — and what that implies for published research.
Chapter 9 — Comparing Groups
ANOVA Visualizer
Adjust group means and within-group variability with sliders and watch the F-statistic and effect size update in real time. Set all means equal to see what happens to \(F\).
Chapters 10 & 11 — Simple and Multiple Regression
Regression Explorer
Add data points, fit the model, and check residuals. Drop in an outlier and watch the slope shift. Used in Chapter 10 for simple regression and again in Chapter 11 to add and remove predictors and see multicollinearity inflate the standard errors.