Who enters? Who gets treatment?
Change recruitment and treatment assignment to separate association, causal effects among volunteers, and population generalization.
Open guided investigation →
Experiment, don't just memorize
Interactive laboratories
The simplest way to understand a mathematical claim is sometimes to try breaking it.
Center & spread
See an extreme observation pull the mean away from the median.
Open laboratory →
Histogram bin explorer
Change the bins without changing a single observation, and inspect the underlying counts.
Open laboratory →
Correlation and outliers
Move one influential point and watch Pearson's r change.
Open laboratory →
Selection bias simulator
Change survey participation while holding everyone's opinions fixed.
Open laboratory →
Central limit theorem
Generate sampling distributions from populations that are anything but bell-shaped.
Open laboratory →
Normal distribution
Change an interval and watch its probability grow or shrink.
Open laboratory →
Bayesian base rates
Change test accuracy and condition prevalence to see why false positives can dominate.
Open laboratory →
Confidence interval coverage
Build fifty intervals repeatedly; see how many contain a fixed true mean.
Open laboratory →
Exact binomial test
Change the head count and read an exact two-sided p-value.
Open laboratory →
Simpson's paradox
Switch between aggregate and stratified published data to discover a reversal.
Open laboratory →
Bootstrap sampling
Resample waiting times with replacement and examine the estimated uncertainty of the mean.
Open laboratory →
Exact randomization test
Enumerate all seventy possible treatment allocations for a tiny randomized experiment.
Open laboratory →
What the line missed
Introduce curvature, refit a line, and watch residual structure emerge.
Open laboratory →
Prefer writing the statistics yourself?The R Workbench runs genuine R through WebAssembly, captures console output and plots, and includes editable examples for bootstrap, Bayes, regression, and hypothesis tests. Runtime loading is optional and starts only when you click Run.
For instructors: predict → experiment → explain teaching prompts and printable student worksheets are available.
The thirteen guided laboratories run entirely in your browser, with no required account or server-side computation.