Teaching
Meet them where they are.
Then move the whole curve.
Teaching is the truest test of understanding.
Featured
Learning by resampling. Open an app and move the numbers.
-
Maximum Likelihood
Pick a model, from binomial to capture-recapture, then watch the likelihood and the log-likelihood peak at the MLE, or sit on a boundary, or go flat when the MLE is not unique.
-
Beta Prior, Binomial Data
Set a beta prior and binomial data, then watch the posterior form between them, with its mean, mode, HPD interval and the beta-binomial predictive for future trials.
-
Blocking
Analyze one experiment as a completely randomized design and as a block design, and see what blocking buys in power and relative efficiency.
-
Step-Down and Step-Up
Sort thousands of p-values and lay the Bonferroni, Holm, Hochberg, Benjamini–Hochberg and Benjamini–Yekutieli lines over them, then count each method's discoveries and how many are false.



