Timeline for Non-parametric bootstrap p-values vs confidence intervals
Current License: CC BY-SA 4.0
4 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jun 22, 2019 at 17:02 | history | bounty ended | CommunityBot | ||
| Jun 17, 2019 at 15:49 | comment | added | EdM | @XavierBourretSicotte I don't think it's quite correct that "bootstrap testing usually involves shifting the datasets to reproduce the null." Each bootstrap sample is an attempt to replicate the original sampling/experiment, using the sample at hand to represent the underlying population. If the statistic in question isn't pivotal, however, then CI developed on the bootstrapped samples won't represent CI developed on the underlying population. So you need to correct the distribution of the statistic toward what it would have been under the null, with BCa or other approaches. | |
| Jun 17, 2019 at 15:31 | comment | added | Xavier Bourret Sicotte | Thanks edm ! If there is a 1-1 rekation between CI and hypothesis test- then why does bootstrap testing usually involve shifting the datasets to reproduce the null? By doing that aren't we getting different results than what we would get by calculating the CI of the distribution of difference for example? | |
| Jun 14, 2019 at 22:25 | history | answered | EdM | CC BY-SA 4.0 |