Exact small-sample resampling
Enumerate a small-sample mean-difference permutation test or the complete bootstrap distribution of a mean.
The example value has been entered. Check it immediately or change the value.
Input is processed in the browser. Refreshing returns a sample value.
Instructions and example
Permutation: fraction of allocations with |mean difference|≥observed. Bootstrap: means of nⁿ ordered samples with replacement, using linear quantiles.
A=1,2 and B=3,4 have difference −2; 2 of 6 allocations are as extreme, giving p=1/3.
Permutation needs at least 2 per group and at most 16 total; bootstrap uses 2–6 values in A and ignores B. Exchangeability and representativeness require review. Floating-point results display at most 12 significant digits.
Frequently asked questions
Does the resampling result vary between runs?
No. All permitted outcomes are enumerated without random draws, so identical inputs give identical results.
Do bootstrap quantiles guarantee population coverage?
No. The 2.5% and 97.5% values are quantiles of the empirical resampling distribution, with substantial small-sample limitations.
Calculation criteria and reference materials
- SciPy: permutation tests
Exact null distributions from reallocating sample labels.
Verification of calculation criteria: · Calculation·verification principles · Report errors