Select a dataset, then run a simulation to see results.
Simulate samples under the null hypothesis many times, recording the mean each time. The histogram is the null distribution; the p-value is the share of simulations at least as extreme as your observed mean. This is a test — there is no confidence interval.