Timeline for ANOVA for comparing same linear models
Current License: CC BY-SA 4.0
4 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 19, 2021 at 16:53 | vote | accept | dm_reader | ||
| May 19, 2021 at 16:43 | comment | added | EdM |
@dm_reader yes; if the appropriate assumptions hold, the p-value is the fraction of experiments in which you would get at least that big a difference in the observed result even if there wasn't a true difference. But you shouldn't be focused on p-values in that way. See this page for example. It's better to pre-specify a level of "significance" and stick with it. Or report the effect and confidence intervals for country.
|
|
| May 19, 2021 at 12:45 | comment | added | dm_reader | thanks for the feedback! If I do it the way you are suggesting, and get for example a p-value of 0.06, does the analysis stay the same? I mean, can I claim for example a 6% significance level? | |
| May 19, 2021 at 12:28 | history | answered | EdM | CC BY-SA 4.0 |