Timeline for Rule of Thumb for Mixed Model
Current License: CC BY-SA 4.0
8 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 1 at 14:25 | vote | accept | 温泽海 | ||
| Jul 31 at 10:56 | history | edited | Frans Rodenburg | CC BY-SA 4.0 |
language-specific code block
|
| Jul 31 at 7:23 | history | edited | Frans Rodenburg |
edited tags
|
|
| Jul 31 at 7:23 | answer | added | Frans Rodenburg | timeline score: 1 | |
| Jul 30 at 21:15 | comment | added | PBulls | It depends on the correlation between within-cluster observations. One rule of thumb I've seen is similar at $n/10$, but $n$ ranges from the total number of observations (low to no correlation) to the number of clusters (very high correlation). | |
| Jul 30 at 19:12 | comment | added | 温泽海 | Yeah I believe it would be more complicated. But is there any at all? | |
| Jul 30 at 19:10 | comment | added | Frans Rodenburg | For mixed models, such a rule of thumb could never be as simple as just a function of the number of parameters, because the effective sample size depends not only on the number of patients but also the variance between patients compared to the variance of the error term $\epsilon$. | |
| Jul 30 at 18:51 | history | asked | 温泽海 | CC BY-SA 4.0 |