Nonparametric Statistics: Rank Tests, Spearman & Bootstrap
Ranks and ties, Mann-Whitney U, Wilcoxon signed-rank, Kruskal-Wallis, Spearman, the sign test and a bootstrap interval: 8 hands-on labs with a food-delivery start-up.
The first lesson is free — no account needed.
About this course
A nonparametric statistics course for data that is skewed, ranked or full of outliers, worked through on one food-delivery start-up's numbers in a spreadsheet. You measure skewness and compare the mean with the median to see when a t-test is a poor fit, rank data with RANK.EQ and RANK.AVG and handle ties, run a Mann-Whitney U test on two routing apps with the normal approximation, test paired before-and-after times with the Wilcoxon signed-rank test, compare three delivery zones with Kruskal-Wallis and CHISQ.DIST.RT, set Spearman rank correlation against Pearson when one outlier distorts the picture, run an exact sign test with BINOM.DIST and build a bootstrap percentile interval from resampled means. For students, analysts and researchers who need the right test without a statistics package. The first lab is free.
Lessons
- When to Use Rank Tests: Skewness, Outliers & the MedianFree
- Ranking Data: RANK.EQ, RANK.AVG and Ties Pro
- Mann-Whitney U Test: Two Independent Groups Pro
- Wilcoxon Signed-Rank Test: Paired Before/After Data Pro
- Kruskal-Wallis H Test: Three or More Groups Pro
- Spearman Rank Correlation vs Pearson Pro
- The Sign Test: Only Up or Down Pro
- Bootstrap Confidence Interval from Resampled Means Pro