Bootstrap confidence intervals, percentile versus better
You have run the bootstrap and hold replicate estimates of a statistic.
How do you build a confidence interval from them, and why can the simple percentile method under-cover?
Show a hint
The percentile method reads quantiles straight off the replicates. What does that quietly assume about the shape of the estimator's sampling distribution?
Your answer
This one is open-ended. Work it through, then check your reasoning against the full solution.