Timeline for confidence intervals for proportions containing a theoretically impossible value (zero)
Current License: CC BY-SA 4.0
17 events
when toggle format | what | by | license | comment | |
---|---|---|---|---|---|
Jul 4 at 16:55 | history | edited | User1865345 | CC BY-SA 4.0 |
deleted 1 character in body
|
Jul 4 at 4:49 | vote | accept | Coris | ||
Jul 3 at 17:31 | history | became hot network question | |||
Jul 3 at 15:39 | answer | added | Sextus Empiricus | timeline score: 4 | |
Jul 3 at 13:28 | history | edited | Coris | CC BY-SA 4.0 |
example in R, in case my explanation isn't clear
|
Jul 3 at 13:19 | history | edited | Coris |
it's about estimation
|
|
Jul 3 at 13:04 | comment | added | Coris | about "Should 0.000001 be included, or excluded as well? " I guess I could put a lower bound by counting the number of individual cases studied in qualitative studies, and dividing that by the population size. However this is assuming we know the population size (and also assuming that there are actually qualitative studies or other similar prior info. But again, all of this is hypothetical, so I'm interested in other possible scenarios too, if it doesn't make the scope of the question too broad). | |
Jul 3 at 12:55 | history | edited | Coris | CC BY-SA 4.0 |
better title I guess
|
Jul 3 at 12:37 | answer | added | Guillaume Dehaene | timeline score: 6 | |
Jul 3 at 10:51 | answer | added | Björn | timeline score: 5 | |
Jul 3 at 9:52 | comment | added | Coris | @SextusEmpiricus About your first comment: yes, that's what I have in mind. | |
Jul 3 at 9:51 | comment | added | Coris | @SextusEmpiricus Thanks, I didn't see this post! As it's a hypothetical question, I don't really have more information than that to share. So I guess (correct me if I'm wrong) that the answer is really on a case-by-case basis, and I should rather ask this question if one day I encounter the issue in a real-life situation. | |
Jul 3 at 9:46 | comment | added | Sextus Empiricus | "compute an estimation that does not include 0 in the first place" This is a bit tricky. If you don't want to include 0, where else do you draw the boundary. Should 0.000001 be included, or excluded as well? The approach to the problem depends on the information that you exactly have. | |
Jul 3 at 9:43 | comment | added | Sextus Empiricus | There are several questions about estimation of a proportion when the observed number is zero. For example Revisiting the Rule of Three. | |
Jul 3 at 9:40 | comment | added | Sextus Empiricus | „I fail to identify any of this "something" in my sample, so the confidence interval includes 0,” This conclusion with 'so' is not so clear and might be clarified better. If I am guessing then it sounds like you are estimating the probability for a binary value when this probability is very small. Is that right? | |
S Jul 3 at 9:31 | review | First questions | |||
Jul 3 at 11:19 | |||||
S Jul 3 at 9:31 | history | asked | Coris | CC BY-SA 4.0 |