Timeline for p-values for hypothesis testing
Current License: CC BY-SA 4.0
7 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 12 at 7:05 | history | edited | Sextus Empiricus | CC BY-SA 4.0 |
added 1222 characters in body
|
| Nov 12 at 5:54 | comment | added | Buck Thorn | This is what I thought. When the pdf is not discrete it is nonsensical to report a probability associated with observing any particular value (an integral might extend over the "measurement uncertainty" but this does not resolve the dilemma). | |
| Nov 10 at 17:59 | history | edited | Sextus Empiricus | CC BY-SA 4.0 |
added 642 characters in body
|
| Nov 10 at 17:33 | comment | added | whuber♦ | Okay--but it is difficult to see how that relates to what you have posted, which discusses how "very small values" can be "compared in a ... relative sense." That sounds like a completely different topic. | |
| Nov 10 at 17:32 | comment | added | Sextus Empiricus | @whuber I read it as: „ why do we use the (integral over) probabilities of all worst values $\int_{x}^\infty f(y) dy$ instead of only the observed single value $f(x)$?“ | |
| Nov 10 at 17:29 | comment | added | whuber♦ | Are you sure you are answering the question? I read it as asking about the "or more extreme" criterion in calculating p-values in hypothesis tests, not about whether numbers are "small" in any sense. | |
| Nov 10 at 17:08 | history | answered | Sextus Empiricus | CC BY-SA 4.0 |