Timeline for Bootstrap vs Wilson score confidence interval
Current License: CC BY-SA 4.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Mar 21, 2021 at 13:48 | history | edited | kjetil b halvorsen♦ | CC BY-SA 4.0 |
added 21 characters in body
|
| Feb 14, 2021 at 13:32 | comment | added | Gideon Kogan | I am familiar with the results that you have demonstrated. Nevertheless, the comments above suggest avoiding bootstrapping and answer my question directly, where your answer demonstrates the comparison between the estimation methods rather than comparing the bootstrap with the closed-form estimation methods. | |
| Feb 14, 2021 at 12:12 | comment | added | R. Cox | In the case of a 95% CI, by 'best coverage' I mean that the coverage is clustered the closest to 95%. | |
| Feb 14, 2021 at 12:02 | comment | added | R. Cox | Good question @Gideon. A Confidence Interval (CI) is a range of values for an unknown parameter with a specified, nominal probability that it contains a feature of interest. This probability is otherwise known as the Confidence Level (CL). For example, a 95% CI predicts with 95% CL that p lies within the CI. The probability that the CI contains this feature of interest is called the Coverage Probability (CP). CP would ideally equal CL and the CP is used as a measure of the quality of methods used for the construction of CIs. So, in the case of a 95% CI, by 'best coverage' I mean 95%. | |
| Feb 13, 2021 at 15:59 | comment | added | Gideon Kogan | What you mean by best coverage? | |
| Feb 12, 2021 at 13:43 | history | answered | R. Cox | CC BY-SA 4.0 |