Timeline for What is the post-hoc power in my experiment? How to calculate this?
Current License: CC BY-SA 4.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Mar 21, 2020 at 11:43 | vote | accept | Blue Various | ||
| Oct 18, 2019 at 14:50 | comment | added | EdM | @BlueVarious p-values map to effect sizes at fixed sample size and test type. If sample size doesn't matter in the original test (e.g., Z-test) neither does the 1:1 relationship between p-value and "post-hoc power"; Hoenig and Heisey show a graph for 1-sided Z-tests. For tests where sample size matters (t-tests and F-tests) Russ Lenth has corresponding tables and formulas here. How to handle unequal variances or sample sizes is inherent in the tests themselves and has nothing extra to do with power calculations. | |
| Oct 11, 2019 at 15:12 | comment | added | Jeremy Miles | In post hoc power, you know the effect size and sample size. They don't change. Hence there is a 1:1 correspondence. | |
| Oct 11, 2019 at 3:31 | comment | added | Blue Various | I know that there is no "correct" post-hoc analysis, as it is often screamed in mass-produced editorial. “Correct post-hoc analysis” I said is synonymous with “post-hoc analysis that many people criticize.” I added it to the main body with emphasis because it was bad. I want to hear about why there is one-to-one correspondence between p-value and power even though there is no formula or code. Although it seems to be one-to-one correspondence, it seems to depend on the effect size and sample size ... | |
| Oct 11, 2019 at 3:30 | comment | added | Blue Various | Thankyou for your commment. I'm sorry, but as I said in the mainbody, verbal explanations are not welcomed. It can be found by googled editorial. Please show me formulas or codes instead of words. Please chunk down words into the formula. | |
| Oct 7, 2019 at 16:29 | history | answered | EdM | CC BY-SA 4.0 |