Two-Sample Permutation Test For Whether Populations Have Different Means

When values are randomly shuffled between two samples, how much does this change the difference between their means?
Sample A

Sample B

Item count: A=20, B=20
Means: A=68.93, B=75.21

meanA - meanB = -6.28

Values for A-B difference of means for 10,000 sample permutations -15 -10 -5 0 5 10 15 0 100 200 300 400 = Difference between the means of the original A and B samples Lower differences: 7.0% of resamples Higher differences: 93.0% of resamples
The graph is divided into two parts that show the percentage of shuffled cases whose difference of means is on each side of the A-B difference for the original values.

open-source site  --  Problems/suggestions?  Notify
(use this link instead of the Back button)