Timeline for When is it important for a practitioner to understand CIs?
Current License: CC BY-SA 4.0
31 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 10, 2024 at 15:46 | comment | added | Sextus Empiricus | A good thought experiment to test/determine whether something is a probability statement is to imagine whether one can gamble on the statement and whether one can make computations about the outcome of the gamble. –– On a related note, for a while I used to participate in a weekly lotery and used a solipsistic Copenhagen interpretation of quantum mechanics by not looking at the outcome of the before last lotery, such that I was always a millionaire , at least in one of the quantum states. –– Just because it has happened and has become deterministic doesn't mean that probability doesn't apply. | |
| Apr 17, 2024 at 17:09 | comment | added | AdamO | It's almost guaranteed that a practitioner is going to read published research which reports a CI. So the answer may be that it's always important to understand CI. The nuance might be at what depth? And in fact this is in most answers you've already received. | |
| Apr 17, 2024 at 8:45 | answer | added | Frans Rodenburg | timeline score: 3 | |
| Apr 17, 2024 at 7:59 | comment | added | Frans Rodenburg | Not sure if it is appropriate to share here, but I made a video about just this to hopefully make it more intuitive: youtu.be/jrUrjv_yM0M | |
| S Apr 17, 2024 at 3:49 | history | suggested | Graham Bornholt | The frequentist tag should lead to better answers | |
| Apr 17, 2024 at 3:43 | review | Suggested edits | |||
| S Apr 17, 2024 at 3:49 | |||||
| Apr 17, 2024 at 3:09 | answer | added | Acccumulation | timeline score: 1 | |
| Apr 16, 2024 at 16:32 | answer | added | EngrStudent | timeline score: 2 | |
| Apr 16, 2024 at 15:14 | answer | added | Durden | timeline score: 4 | |
| Apr 16, 2024 at 6:01 | comment | added | Graham Bornholt | @whuber Thanks. I see now that my earlier comment misunderstood the meaning of your comment on the answer. | |
| Apr 16, 2024 at 3:40 | answer | added | jonas-eschle | timeline score: 2 | |
| S Apr 16, 2024 at 2:41 | history | suggested | Peter Mortensen | CC BY-SA 4.0 | Copy edited (e.g. ref. <https://en.wiktionary.org/wiki/there#Pronoun>). Removed unnecessary formatting. |
| Apr 16, 2024 at 1:40 | review | Suggested edits | |||
| S Apr 16, 2024 at 2:41 | |||||
| Apr 15, 2024 at 20:45 | answer | added | Michael Lew | timeline score: -1 | |
| Apr 15, 2024 at 14:03 | answer | added | Björn | timeline score: 7 | |
| Apr 15, 2024 at 13:05 | answer | added | Tasos Papastylianou | timeline score: -1 | |
| Apr 15, 2024 at 12:13 | comment | added | whuber♦ | @Graham The Bayes/Frequentist/Whatever approach is separate from this. If we cannot make probability statements about completed events, then we have destroyed all possibility of modeling data probabilistically. Such confusion of temporal sequence, knowledge, and probability models is counterproductive. | |
| Apr 15, 2024 at 12:11 | comment | added | whuber♦ | @Yair The "probability statement" is being made about the interval, not the parameter. | |
| Apr 15, 2024 at 11:20 | comment | added | Graham Bornholt | @StephanKolassa I think what whuber is alluding to the fact that "no probabiity any more" would not be accurate for a Bayesian. [The Bayesian posterior probability for an observed 99% CI, for example, could be any number between 0% and 100%, it just depends on the choice of prior.] | |
| Apr 15, 2024 at 9:45 | comment | added | Dikran Marsupial | @YairDaon no (non-trivial) frequentist probability statement can be made about them. There are also Bayesian probability statements and trivial frequentist probability statements, 0 and 1, but at that point they are more observations/knowledge than probabilities. Also relevant Q stats.stackexchange.com/questions/26450/… | |
| Apr 15, 2024 at 9:42 | comment | added | Dikran Marsupial | Whenever they use them. Nobody should be using statistics they don't understand (or at least they should be collaborating with someone that does understand them). | |
| Apr 15, 2024 at 9:34 | comment | added | Yair Daon | @whuber AFAIK Neymann himself took the view I cite above. See accendoreliability.com/…: "...The parameter is an unknown constant and no probability statement concerning its value may be made...". | |
| Apr 15, 2024 at 7:26 | comment | added | Stephan Kolassa | @whuber: I don't quite follow your logic. The CI either contains the parameter or doesn't, and in this sense there "is no probability" any more. But we still don't know whether it does. | |
| Apr 15, 2024 at 3:13 | comment | added | whuber♦ | Re "no probability any more:" that is not so. If it were correct, then one should be able either to state definitely whether the interval covers the parameter. | |
| Apr 15, 2024 at 1:46 | answer | added | Harvey Motulsky | timeline score: 7 | |
| Apr 15, 2024 at 1:43 | history | became hot network question | |||
| Apr 14, 2024 at 23:07 | answer | added | jginestet | timeline score: -2 | |
| Apr 14, 2024 at 19:28 | comment | added | kjetil b halvorsen♦ | Some similar Qs: The importance of a correct interpretation of a confidence interval, What's wrong with this interpretation of a 95% confidence interval?, What, precisely, is a confidence interval? | |
| Apr 14, 2024 at 19:23 | history | edited | kjetil b halvorsen♦ | edited tags | |
| Apr 14, 2024 at 18:00 | answer | added | Demetri Pananos | timeline score: 14 | |
| Apr 14, 2024 at 17:41 | history | asked | Yair Daon | CC BY-SA 4.0 |